Ownership Structure and Executive Compensation in Canadian ... · Ownership Structure and Executive...
Transcript of Ownership Structure and Executive Compensation in Canadian ... · Ownership Structure and Executive...
Ownership Structure and Executive Compensation in Canadian Corporations
A Thesis Submitted to the College of
Graduate Studies and Research
In Partial Fulfillment of the Requirements
For the Degree of Master of Science in Finance
In the Department of Finance and Management Science
Edwards School of Business
University of Saskatchewan
Saskatoon, Saskatchewan, Canada
By
WEIWEI JIANG
Copyright Weiwei Jiang, April, 2011. All rights reserved.
i
Permission to Use
In presenting this thesis in partial fulfilment of the requirements for a Postgraduate
degree from the University of Saskatchewan, I agree that the Libraries of this University may
make it freely available for inspection. I further agree that permission for copying of this thesis
in any manner, in whole or in part, for scholarly purposes may be granted by the professor or
professors who supervised my thesis work or, in their absence, by the Head of the Department or
the Dean of the College in which my thesis work was done. It is understood that any copying or
publication or use of this thesis or parts thereof for financial gain shall not be allowed without
my written permission. It is also understood that due recognition shall be given to me and to the
University of Saskatchewan in any scholarly use which may be made of any material in my
thesis.
Requests for permission to copy or to make other use of material in this thesis in whole or
part should be addressed to:
Head of the Department of Finance and Management Science
Edwards School of Business
University of Saskatchewan
25 Campus Drive
Saskatoon, Saskatchewan S7N 5A7
ii
Abstract
Agency theory, proposed by previous studies such as Guidry, Leone, and Rock (1999)
and Arya and Huey-Lian (2004), suggests that bonus and other accounting-metric-based
compensation can motivate managers to perform well in the short horizon while equity-based
compensation, such as restricted shares and stock options, can serve the purpose of aligning the
long run interests of shareholders and managers. The empirical evidence, for example Jensen and
Murphy (1990), Kaplan (1994), Hall and Liebman (1998), Murphy (1999), Zhou (2000), and
Chowdhury and Wang (2009), confirms that incentive compensation is popular in many
countries. However, recent studies suggest that the relation between performance and incentive
compensation is weak. Shaw and Zhang (2010) find that CEO bonus compensation is less
sensitive to poor earnings performance than it is to good earnings performance. Fahlenbrach and
Stulz (2011) study the relation between bank performance during the 2008 bank crisis and the
bonus and equity-based compensation of bank CEOs. They find that banks with CEOs whose
incentives were better aligned with the interests of shareholders performed worse than other
banks.
This study examines whether ownership structure can explain the differences among
compensation structures of chief executive officers (CEOs). In particular, we examine the
compensation structure of three distinct groups: family-controlled, institution-controlled, and
widely-held firms. We distinguish these three kinds of firms to represent different levels of
market imperfection. Compared with family-controlled and institution-controlled firms, widely
held firms have dispersed ownership. The most significant weakness of a widely-held ownership
structure is the lack of shareholder monitoring due to the unmatched benefit and cost of
monitoring for small shareholders. In contrast, a holder of a large block of shares will have the
same monitoring costs but the benefits to this shareholder from monitoring management and
reducing agency costs would be substantial and larger than the costs of monitoring. Thus the
presence of a large shareholder will reduce the agency costs. In addition, large shareholders may
be willing to spend time and effort continuously to collect more information on management
performance or to estimate the firm’s investment projects. This behaviour will reduce the
problems that arise from information asymmetry and will decrease the waste of free cash flows
by managers.
iii
Both family-controlled firms and institution-controlled firms have large shareholders.
However, whether or not the control shareholders are playing an active monitoring role is still an
important issue. From the viewpoint of aligning the interests of managers and shareholders, the
family-controlled group is superior to the institution-controlled group. First, institutions are more
flexible in moving their ownership from one firm to another depending on performance. If the
costs of monitoring are high in comparison to the costs of rebalancing portfolios, institutions will
choose to rebalance instead of monitoring. In contrast, a family that controls a firm does not have
this flexibility. Second, family-controlled firms generally assign influential positions to family
members whose focus is in line with that of the family group. Even though a non family member
may be appointed as the manager, the level of monitoring is significant given the high ownership
concentration by the family. However, the level of monitoring by a family may not necessarily
translate into a reduction of agency costs for minority shareholders. Indeed, previous studies
suggest that significant family ownership may lead to agency costs of its own. The family may
divert company resources for its own benefit despite the presence of a manager who may or may
not be a family member. Essentially, the family and the manager can collude to spend on perks
and personal benefits at the expense of minority shareholders. Chourou (2010) suggests that
excessive compensation of chief executive officers at some family owned Canadian corporations
may be viewed as expropriation of minority rights.
Overall, the main objective of this study is to examine whether block-holder monitoring
is a substitute to the incentive components of compensation. We propose that as we move from
widely-held to institution-controlled the level of monitoring may or may not increase. However,
as we move further into higher control, as may be suggested by family ownership, the level of
monitoring will increase but this monitoring may not necessarily reduce agency costs. The
results show that the institution-controlled firms pay significantly less bonus compensation per
dollar of assets than widely-held firms but the differences in equity based compensation are not
significant. In addition, the family-controlled corporations offer the lowest performance-based
compensation, bonus per dollar of assets, in comparison to the institution-controlled and the
widely-held groups. These results indicate that the family-controlled Canadian corporations rely
more on monitoring managers than paying them incentive payments in the form of bonus
payments. In addition, our results indicate that the institutions which control corporations may be
monitoring the managers of these corporations but this monitoring does not significantly reduce
iv
the need for the long-term incentive components of compensation. This result suggests that
institutions may monitor the short-term performance effectively but they may prefer rebalancing
their portfolio rather than monitoring long term performance.
v
ACKNOWLEDGMENTS
My profound appreciation and gratitude goes to my supervisors Professor George
Tannous and Dr. Fan Yang first for being an admirable coach in guiding me through my thesis. I
sincerely acknowledge their rigorous commitment and endless effort to ensure the quality of this
thesis. They in depth knowledge, intellectual ability and generosity have helped me to have a
good understanding of the Corporate Governance issues. Without the large amount of time and
energy on helping me from them, this thesis could never have been completed. It is my pleasure,
honor, and luck to have such a great opportunity to learn from them. I also would like to dedicate
sincere appreciation to my supervisory committee member Dr. Zhuyu Wu, as well as my external
examiner, Dr. Don Cyr, for their insight comments and suggestions.
I am thankful to all my professors to their contributions to my knowledge in Finance. I
am thankful to the countless help and support that I received from Dr. Marie Racine. I am
thankful to Dr. Abdullah Mamun, Dr. Craig Wilson, and Dr. Dev Mishra for their contributions
to my knowledge in Finance. I am thankful to Ms. Brenda Orischuk, for her assistance during my
studies at the University of Saskatchewan. I acknowledge the encouragements and help of my
friends and classmates. Special thanks to Yuting Fu, Eva Yang, Mo Zhou, and Lucy Li.
Finally, my deepest love and gratitude go to my best parents and my husband for their
unconditional love, encourage, and understanding.
vi
TABLE OF CONTENTS
TABLE OF CONTENTS ..................................................................................................... i ABSTRACT ........................................................................................................................ ii ACKONWLEDGEMENTS ................................................................................................. v LIST OF FIGURES ......................................................................................................... viii LIST OF TABLES .............................................................................................................. ix CHAPTER 1 INTRODUCTION ............................................................................................................... 1 CHAPTER 2 LITERATURE REVIEW ....................................................................................................6
2.1 The Components of CEO Compensation .................................................................. 6 2.2 Agency Theory and Incentive Compensation ........................................................... 7 2.3 Empirical Evidence Regarding the Relation between Pay and Performance ........... 8 2.4 Ownership Structure .............................................................................................. .10
2.4.1 Widely-held Group Versus the Concentrated Group………………………....12 2.4.1 Family-controlled Group Versus Institution-controlled Group……………....14
2.5 Other Factors ........................................................................................................... 17
CHAPTER 3 THEORETICAL ARGUMENTS AND HYPOTHESES ..................................................19 CHAPTER 4 DATA ................................................................................................................................23
4.1 Ownership Structure ............................................................................................... 23 4.2 CEO Compensation ................................................................................................ 24 4.3 Variables ................................................................................................................ 27
4.3.1 Dependent Variables……………………………………………………….....27 4.3.2 Control Variables……………………………………………..……………... 27 4.3.3 DummyVariables……………………………………………..……………... 28
CHAPTER 5 DESCRIPTIVE STATISTICS AND UNIVARIATE TESTS ...........................................30
5.1 Permanment CEO Firms ......................................................................................... 30 5.2 Transient CEO Group ............................................................................................. 32 5.3 Aggregate Sample ................................................................................................... 35
5.3.1 Descriptive Statistics....……………………………………………………....35 5.3.2 Univarite Variables……………………………………………..…………....36
CHAPTER 6 MULTIVARIATE ANALYSIS.........................................................................................39
6.1 Natural Logarithm of Compensation as Dependent Variable ................................. 41 6.2 Bonus and Incentive Compensation as Percentage of Total Assets ....................... 43
6.2.1 Annual Bonus as a Percentage of Total Assets…………………………….....44
vii
6.2.2 Contingent Compensation as a Percentage of Total Assets…………….….....45 6.3 Bonus and Incentive Compensation as Percentage of Total Assets ....................... 46
6.3.1 Annual Bonus as a Percentage of Total Pay……………………………….....46 6.3.2 Contingent Compensation as a Percentage of Total Pay…………………......47
6.4 The Relation between Compensation and Total Market Return ............................. 48
CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER RESEARCH ......... 51
REFERENCES ................................................................................................................. 55
viii
LIST OF FIGURES
Figure 1.1: Sarlay in monetary terms paid by permanent CEO firms .................................90
Figure 1.2: Bonus in monetary terms paid by permanent CEO firms ..................................90
Figure 1.3: Contingent compensation in monetary terms paid by permanent CEO firms ....91
Figure 1.3: Total compensation in monetary terms paid by permanent CEO firms .............91
Figure 2.1: Salary as a percentage of total compensation paid by permanent CEO firms ...92
Figure 2.2: Annual Bonus as a percentage of total compensation paid by permanent CEO firms..............................................................................................................................92
Figure 2.3: Contingent Compensation as a percentage of total compensation paid by permanent CEO firms ............................................................................................................93
Figure 3.1: Changes in compensation following CEO turnovers in family-controlled firms..............................................................................................................................94
Figure 3.2: Changes in compensation following CEO turnovers in institution-controlled firms..............................................................................................................................94
Figure 3.3: Changes in compensation following CEO turnovers in widely-held firms ........95
Figure 4.1: Changes in compensation following CEO retirements in family-controlled firms..............................................................................................................................96
Figure 4.2: Changes in compensation following CEO retirements in institution-controlled firms..............................................................................................................................96
Figure 5.1: Total compensation as a function of total assets (all data) .................................97
Figure 5.2: Total compensation as a function of total assets (assets sizes of $20.48 million-9 billion) ..................................................................................................................98
Figure 5.3: Total compensation as a function of total assets (assets sizes of $9.1 billion-$55 billion) ..................................................................................................................99
Figure 5.4: Total compensation as a function of total assets (asset sizes larger than $55 billion)..............................................................................................................................100
Figure 6.1: Salary and Bonus are as a function of total assets (all data) ..............................101
Figure 6.2: Salary and Bonus are as a function of total assets (assets sizes of $20.48 million-55 billion) ..................................................................................................................102
ix
LIST OF TABLES
Table 1.1: Mean value of CEO monetary compensation in permanent CEOs firms .............61
Table 1.2: Mean value of salary, annual bonus, and contingent pay as a percentage of total pay in permanent CEOs firms .........................................................................................61
Table 2.1: Descriptive statistics in the transient CEO group .................................................62
Table 2.2: Comparison of the compensation of incoming CEOs and that of their predecessors..............................................................................................................................63
Table 3: Descriptive statistics related to the dependent and control variables ......................64
Table 4: Descriptive statistics related to the dependent and control variables in three different ownership structures ............................................................................................65
Table 5: Descriptive statistics related to the dependent and control variables in three different industries ..............................................................................................................66
Table 6.1: T-test: two sample assuming unequal variances ...................................................67
Table 6.2: Descriptive statistics mean value annual bonus and contingent pay ....................67
Table 7: Descriptive statistics mean for the control variables ...............................................68
Table 8.1: Correlation of variables in the aggressive sample ................................................69
Table 8.2: Correlation of variables in the family-controlled firms ........................................70
Table 8.3: Correlation of variables in the institution-controlled firms ..................................71
Table 8.4: Correlation of variables in the widely-held firms .................................................72
Table 9.1: The impact of ownership structure on the natural log of annual bonus (OLS) ....73
Table 9.2: Natural log of annual bonus in widely-held, institution-controlled, and family-controlled firms (OLS) .........................................................................................74
Table 10.1: The impact of ownership structure on the natural log of contingent compensation (OLS) ...................................................................................................................75
Table 10.2: Natural log of contingent compensation in widely-held, institution-controlled, and family-controlled firms (OLS) .............................................................................76
Table 11.1: The impact of ownership structure on the annual bonus as a proportion of total assets (Tobit) .................................................................................................................77
x
Table 11.2: Annual bonus as a proportion of total assets in widely-held, institution-controlled, and family-controlled firms (Tobit) .....................................................................78
Table 12.1: The impact of ownership structure on the contingent compensation as a proportion of total assets (Tobit) ................................................................................................79
Table 12.2: Contingent compensation as a proportion of total assets in widely-held, institution-controlled, and family-controlled firms (Tobit) ...................................................80
Table 13.1: The impact of ownership structure on the annual bonus as a proportion of total pay (Tobit) ..................................................................................................................81
Table 13.2: Annual bonus as a proportion of total pay in widely-held, institution-controlled, and family-controlled firms (Tobit) ............................................................................82
Table 14.1: The impact of ownership structure on the contingent compensation as a proportion of total pay (Tobit) ...................................................................................................83
Table 14.2: Contingent pay as a proportion of total pay in widely-held, institution-controlled, and family-controlled firms (Tobit) ............................................................................84
Table 15: Correlation of variables in the family-controlled, institution-controlled, and widely-held firms .............................................................................................................85
Table 16.1: The impact of ownership structure on the natural log of annual bonus (OLS) ..86
Table 16.2: Natural log of annual bonus in widely-held, institution-controlled, and family-controlled firms (OLS) .........................................................................................87
Table 17.1: The impact of ownership structure on the natural log of contingent pay (OLS) 88
Table 17.2: Natural log of contingent compensation in widely-held, institution-controlled, and family-controlled firms (OLS) .............................................................................89
1
CHAPTER 1 Introduction
In theory, efficient pay contracts bond executive compensation with firm performance,
and offer strong incentives for executives to act in shareholders’ best interests. Guidry, Leone,
and Rock (1999) suggest that bonus and other accounting-metric-based compensation can
motivate managers to perform well in the short run. Jenson and Murphy (1990) propose that
CEO ownership of their firm’s stock is the largest CEO performance incentives. Lamber,
Larchker, and Verrecchia (1991) point out that the stock-based compensation can mitigate
agency problems. Similarly, Arya and Huey-Lian (2004) propose that equity-based
compensation, such as restricted shares and stock options, can align the long-term interests of
shareholders and managers.
A number of empirical studies confirm that incentive pay, including both short-term
bonus and long-term equity-based compensation, are used to reduce agency costs. Kaplan (1994)
provides evidence suggesting that the fortunes of Japanese top executives are related to stock
performance and to factors that are conducive to stock and earning performance. Hall and
Liebman (1998) provide evidence suggesting that firm performance is correlated to CEO
compensation. Murphy (1999) reports that the relation between compensation and performance
in the United States is stronger than the same relation in other countries. In the context of Canada,
Zhou (2000) suggests that executive compensation is positively correlated to firm performance
with an overall weak relationship. Another Canadian study, Chowdhury and Wang (2009), find
that contingent pay and its ratio to total pay have been increasing in Canada from 1995 to 2002.
Recently, some studies show that the relation between performance and incentive
compensation is weak. Shaw and Zhang (2010) find that CEO bonus compensation is less
sensitive to poor earnings performance than it is to good earnings performance. They suggest
that CEOs get rewards even with poor firm performance. Similarly, Fahlenbrach and Stulz (2011)
find no evidence to support the proposition that banks with CEOs whose incentives were not
well aligned with the interests of their shareholders performed worse.
This study examines whether ownership structure can explain the differences among the
levels and structures of chief executive officer (CEO) compensation. In particular, we examine
the compensation of three distinct groups: family-controlled, institution-controlled, and widely-
held firms. We distinguish these three kinds of firms to represent different levels control and
2
monitoring provided by widely disbursed and concentrated ownership and by institutions as
opposed to families.
Previous studies suggest that the most significant weakness of a widely-held ownership
structure is the lack of shareholder monitoring due to the unmatched benefit and cost of
monitoring for small shareholders. Demsetz (1983) suggests that when ownership is widely
dispersed across many individuals and institutions, shareholders cannot exercise real power to
oversee managerial performance in modern corporations. The existence of at least one large
shareholder will reduce the agency costs and asymmetric information. McConaughy et al. (1998)
examine the efficiency, measured by sales growth, and value of family-controlled firms. Their
family-controlled firms are defined as public corporations whose CEOs are either the founder or
a member of the founder’s family. Controlling for size, industry and ownership effects, they
apply a matched-pairs methodology. Their key finding is that family controlled firms are more
valuable and efficient than firms of the same size, in the same industry, and with similar
managerial ownership. Their findings also emphasize that who owns the shares is more
important than ownership concentration. Firth, Fung and Rui (2006) indicate that concentrated
ownership reduces agency costs.
Both family-controlled firms and institution-controlled firms have large shareholders.
David, Kochhar, and Levitas (1998) argue that institutions have the obligation to know and
protect what they invest. They should take proactive actions, so that the managements of investee
firms work towards maximizing shareholder value. We propose that firms with a concentrated
ownership structure would behave differently depending on whether they are institution-
controlled or family-controlled. Shleifer and Vishny (1986) show that institution block holders
do not usually interfere with management but they perform better monitoring than small
shareholders in widely-held firms. However, whether or not institutions are effective in
controlling managers is still an important issue. Interviews with six investment managers that
control significant pension assets in Canada reveal that institutions actively communicate with
and monitor managements of the firms they invest in but these investment managers stopped
short of claiming that they attempt to exercise control over managements. Institutions are more
flexible in moving their ownership from one firm to another depending on performance. If the
costs of monitoring are high in comparison to the costs of rebalancing portfolios, institutions will
choose to rebalance instead of trying to change managerial attitudes and decisions. Therefore, we
3
propose that institutions are better positioned than individuals to monitor managers and provide
recommendations but they are not likely to exert significant and effective control over the
managers of the firms in which they have control.
In contrast, a family that holds a controlling portion of voting shares is likely to have a
significant and personal interest in the firm. They family would generally assign influential
positions to family members or to managers who are controlled indirectly by the family. Thus,
the managers of family-controlled firms are likely to be monitored more closely than the
managers of institution-controlled firms and they are more likely to be directed and influenced
by the controlling family. Many studies suggest that the fortunes of the managers of family-
controlled firms are very much tied with the fortunes of the families that control their firms.
Fama and Jensen (1983) and Demsetz (1983) suggest that a manager who cooperates with the
controlling family can guarantee employment at an attractive salary as long as the firm’s
performance is in line with the industry’s performance and the decisions of the manager are
consistent with the expectations of the family. Furthermore, Chen and Kensinger (1988) suggest
that managers of family-controlled firms may avoid risky ventures which might be desirable for
outside shareholders. Morck et al. (2000) suggest that the concentration of family wealth in a
business and the concern over the family legacy may explain why family-controlled firms may
display excessive risk-aversion and forego profitable expansion strategies and mergers. Chourou
(2010) suggests that excessive compensation of chief executive officers at some family owned
Canadian corporations may be a sign of cooperation between the controlling family and the CEO
and can be perceived as expropriation of minority rights. These arguments suggest that the
compensation packages of family-controlled managers are likely to be competitive in the market
for managerial talent, encourage good performance, promote cooperation with the controlling
family, and discourage managers from taking excessive risk. Therefore, we propose that the
incentive compensation in concentrated ownership firms may or may not vary depending on
whether the firm is family-controlled or institution-controlled. However, the relation between
incentive compensation and performance should be stronger when a family is the source of
ownership concentration.
Overall, we propose that incentive compensation may or may not vary across ownership
structures but the relation between incentive compensation and performance is likely to be weak
at the widely-held firms, stronger at institution-controlled firms, and strongest at the family-
4
controlled firms. Furthermore, we propose that if incentive compensation is different across the
ownership structures the differences are likely to be significant with the bonus component. We
suggest that the ownership structure that provides the strongest monitoring is likely to rely more
on bonus compensation as contingent component may reward managers due to factors beyond
their control. Overall, it is easier to see how managerial actions affect the measures upon which
bonuses are based rather than to point out the managerial actions that affect contingent
compensation.
We examine our theory regarding the relationship between ownership structure,
compensation, and monitoring by also considering how incoming CEOs are compensated in
comparison to their predecessors. Previous studies provide little information on the differences.
Ocasio (1994) suggests that CEO compensation is affected by the CEO’s tenure. In particular,
experience in the industry and in similar position may enable the incoming CEO to negotiate a
high compensation package and a structure that is in the best interests of the CEO. On the other
hand, the departure of a CEO could be seen as an opportunity for a firm to re-establish its own
priorities and to design the compensation package to promote the interests of shareholders and
the ultimate power brokers of the firm. Accordingly, we expect that the compensation of the
incoming CEOs would be structured differently than the compensation of their predecessors.
Furthermore, we propose that the structure of the compensation packages of incoming CEOs will
vary depending on the ownership structure.
Finally, we examine the relation between equity performance and compensation.
Previous studies find this relation to be positive but weak. We propose that separating firms
across ownership structures may reveal that the relation is significant for one ownership structure
and not significant for another. In particular, we propose that the relation is not likely to be
significant in the widely-held firms and institution-controlled firms but it is more likely to be
significant for firms in the family-controlled firms. In our view, the strength of the relation
between incentive compensation and equity performance should be inversely related to the level
of monitoring by the owners of the firms.
The remainder of this thesis is organized into seven sections. In Chapter 2, we review the
prior literature on ownership structures, CEO incentive compensation, and firm performance.
Theoretical arguments and hypothesis are discussed in Chapter 3. In Chapter 4, we describe the
data, define the variables, and explain the methods. In Chapter 5, we discuss our descriptive
5
statistics and results of univariate tests. In Chapter 6, we present and analyze the results of
multivariate tests. Chapter 7 covers conclusions and recommendations for future research.
6
CHPATER 2
Literature Review
This study examines whether ownership structure can explain the differences among
compensation structures of CEOs. In particular, we examine the compensation structure of three
distinct groups: family-controlled firms, institution-controlled firms and widely-held firms. Thus,
this chapter divides the literature into four sections: Section 2.1 discusses the components of
incentive pay. Section 2.2 reviews the literature that examines agency theory and incentive pay.
Section 2.3 provides a review of the empirical evidence regarding the relation between pay and
performance. Section 2.4 discusses the literature related to ownership structure and incentive pay,
and Section 2.5 presents some additional factors that may impact incentive pay.
2.1 The Components of CEO Compensation
Although structures of CEO compensation vary across firms and over time, previous
studies, for example Gray and Cannella (1997), find that CEO compensation contracts are
usually structured as a combination of cash compensation (salary and annual bonus) and long-
term compensation (stock-based and option-based compensation). Both the annual bonus and
long-term incentive pay are usually set based on some measures of firm performance. Yet, there
is a major difference between the two. The annual bonus is mostly based on accounting earnings
such as return on equity or return on assets while the long-term incentive pay is based on stock
returns.
Healy (1985) examines the nature of bonus payments. He notes that when actual firm
performance is below some minimum threshold, no funds are allocated to the bonus pool. As
firms perform better than the minimum, funds are linearly related to firm performance. When
firm performance is up to a ceiling, the bonus pool will be capped. He argues that since a large
part of CEO compensation is short-term bonuses based on accounting earnings, managers are
likely to choose to maximize their short-term bonuses.
Murphy (1999) reports that every profit-oriented company provides performance-based
bonus payments paid annually in cash. His findings suggest that if the CEO meets the
performance target, the CEO will receive the bonus which is usually determined as a given
percentage of his/her salary. Murphy (1999) finds that bonus contracts are usually written based
on accounting earnings and not explicitly on stock returns. Yet, he argues that annual bonus
7
payments have the potential of strengthening the alignment of interests between managers and
shareholders as they link the annual incentive awards to the future value of common shares.
Guidry, Leone, and Rock (1999) and Arya and Huey-Lian (2004), suggests that bonus and other
accounting-metric-based compensation can motivate the managers to perform well in the short
horizon while equity-based compensation, such as restricted shares and stock options, can serve
the purpose of aligning the long run interests of shareholders and managers.
2.2 Agency Theory and Incentive Compensation
Jensen and Meckling (1976) argue that an agency relationship arises when principals
appoint agents to make and execute decisions on behalf of the principals. Both the principal and
the agent want to maximize their respective utilities. Therefore, the agent will focus on
maximizing his/her own utility, not that of the principal. Unless the interests of the principal and
the agent are the same, the decisions of the agent will lead to suboptimal results from the
perspective of the agent. Hence, the principal would like to provide the agent enough incentives
to ensure that the agent acts in the best interests of the principal. The problem of the principal is
to determine the optimal incentive package that does not offer excessive incentives to the agent.
Furthermore, alignment of the interest between the agent and the principal cannot be
achieved at zero cost. Jensen and Meckling (1976) define agency costs to be consisting of three
components: the monitoring expenditures by the principal, the bonding expenditures by the agent,
and the residual loss from the suboptimal decisions. They argue that managerial ownership in the
firm would reduce the conflicts between management and shareholders because managers would
pay a share of the agency costs proportional to their ownership. Thus, Jensen and Meckling
(1976) suggest that management ownership is a good way to align the interests of managers and
shareholders and to reduce agency costs.
These arguments suggest that a mechanism for reducing agency problems between
managers and owners is the employment contract which specifies compensation and its
components. Shavell (1979) proposes that different forms of compensation have different
incentive effects on CEOs. Lambert, Larcker, and Verrecchia (1991) suggest that if a manager’s
compensation is tied to the stock price, the agency problem which includes overly short-sighted
behaviours can be mitigated. Gray and Cannella (1997) argue that incentive pay is related more
with long-term performance and value appreciation of firms and non-incentive compensation
8
aims at providing the CEOs with a stable stream of cash flows. Daily, Johnson, Ellstrand, and
Dalton (1998) propose that incentive pay and non-incentive pay induce different levels of risk
and incentive objectives. Arya and Huey-Lian (2004) suggest that equity-based compensation,
such as restricted shares and stock options, can serve the purpose of aligning the long run
interests of shareholders and managers. Chowdhury and Wang (2009) argue that companies are
structuring CEOs compensation as a combination of incentive pay and non-incentive pay to align
the interests of owners and CEOs.
Another mechanism for reducing agency problems between managers and owners is the
ownership structure. Fama and Jensen (1983) argue that concentrated ownership by outsiders has
the same effects as managerial ownership in reducing agency costs. Thus, the presence of a
shareholder that owns a significant portion of voting rights can be a substitute to significant
managerial ownership in large corporations, large professional partnerships, and mutual
companies. They indicate that concentrated shareholdings by outsiders create more effective
monitoring of managers, which can improve firm performance. Ang, Cole, and Lin (2002)
empirically examine how agency costs vary with a firm’s ownership structure using a sample of
1,708 small US corporations. Their results support the theories of Jensen and Meckling (1976)
and Fama and Jensen (1983) about ownership structure and the alignment of interests between
managers and shareholders. In particular, they find that when an outsider manages the firm,
agency costs are higher. Also, agency costs vary inversely with the manager’s ownership level.
Moreover, agency costs are positively related to the number of non-manager shareholders.
2.3 Empirical evidence regarding the relation between pay and performance
Section 2.2 suggests that incentive compensation is the tool to align the interests of
managers and owners. The empirical evidence confirms that incentive compensation is popular
in many countries and is widely accepted by companies. For example, Barenbaum and Schubert
(1993) find that in 1988 more than 90% of the 400 largest industrial and service companies in the
United States (US) used stock options as part of their compensation packages. However, recent
studies suggest that the relation between performance and incentive compensation is weak.
Jensen and Murphy (1990) empirically examine the relationship between executive
incentives and performance by using over 2000 CEOs data. Their results show that a 10%
change in firm value leads to 0.33% change in total CEO compensation. These results
9
demonstrate that CEO wealth is not significantly related to shareholder wealth. They argue that
CEO ownership of their firm’s stock is the largest CEO performance incentive. However, the
holdings of CEOs are small and decreasing. Leonard (1990) finds results different from Jensen
and Murphy (1990). The author indicates that corporate success does not have an impact on the
level of executive pay using 439 large US corporations over a period of 1981-1985. Kaplan
(1994) indicates that incentive pay which is tied to stock performance and to factors that are
conducive to earnings performance affects the fortunes of Japanese top executives. He also finds
that the stock performance is less related to the fortunes of Japanese managers than those of US
managers. A related study of Hall and Liebman (1998) document that firm performance is
strongly correlated to CEO compensation using data over the period 1980-1994. In particular,
they show that salary and bonus are weakly related to firm performance. However, in
comparison to salary and bonus, the equity-based pay works better to align the interests of CEOs
and shareholders. Similarly, Murphy (1999) uses a number of control variables to clarify the pay-
performance relationship. By using US data covering 1970 to 1996, and international data in
1997, he offers the following important insights. First, in larger firms, if the levels of pay are
high, the pay-performance sensitivities are low. Second, the relationship between the level of pay
and pay-performance sensitivity are more significant in industrial firms than in utilities. Third, in
the US the level of pay for performance is much higher than the pay for performance in other
countries. Fourth, although the incentive pay-performance relations are significant, managers
should not be left alone to design performance-based compensation.
More recent, many studies have shown significant positive relation between pay and
performance, but with rather weak pay-performance sensitivity (Jeppson, Smith, and Stone 2009;
Shaw and Zhang 2010; Fahlenbrach and Stulz 2011). Jeppson, Smith, and Stone (2009) examine
the relationship between CEO compensation and several measures of firm performance using
200 large public companies in 2007 which filed proxy statements with the SEC. They do not find
a strong relationship between CEO compensation and firm performance. The exception is total
revenue, but with a low R2. They also find that CEO compensation is positively related to firm
size. Similarly, Shaw and Zhang (2010) find that CEO bonus compensation is less sensitive to
poor earnings performance than it is to good earnings performance using data over the period
1992-2005. They find no evidence supporting that CEOs are punished for poor firm performance.
Indeed, CEOs are even rewarded with poor performance. Similarly, Fahlenbrach and Stulz (2011)
10
study the relation between bank performance during the 2008 bank crisis and the bonus and
equity-based compensation of bank CEOs. They find that banks with CEOs whose incentives
were better aligned with the interests of their shareholders performed worse. Their results show
that both cash bonus and stock options do not have an adverse impact on bank performance
during the crisis.
There are two studies that consider that relation between incentive pay and performance
in Canadian companies. Zhou (2000) considers executive compensation over the period 1991-
1995 inclusive and provides several insights. The results show that executive compensation is
positively correlated to firm performance but the overall relationship is weak. The author finds
that firm size has a positive impact on CEO compensation. Moreover, smaller firms exhibit a
strong negative correlation between the probability of CEO turnover and stock performance. We
extend their study by examining contingent compensation using more recent observations that
cover 5 years instead of 3 and we classify companies on the basis of their ownership structure.
Another Canadian study that may have objectives similar to those of this study is
Chowdhury and Wang (2009). Using data related to the TSE 300 firms from 1995 to 2002, they
find that contingent pay in Canada, both in monetary terms and as percentage of total pay, has
been increasing during the study period. They argue that this finding is consistent with an
implication of agency theory. Namely, boards of directors seem to be trying to raise CEO
contingent pay to ensure better performance. Their results show that firm size and investment
opportunities positively affect contingent pay. However, the key limitation in Chowdhury and
Wang (2009) is that they investigate CEO contingent compensation only in institution-controlled
firms. We extend their work by examining family-controlled as well as institution-controlled
firms.
2.4 Ownership Structure
This study examines whether ownership structure can explain the differences among the
compensation structures of chief executive officers (CEOs). Therefore, in this section we review
the findings of previous studies regarding the impact of ownership structure on corporations.
During the era of Berle and Means (1932), the theory of the firm was developed under
the assumption that organizations have widely-held ownership. La Porta et al (1999) suggest that
the widely-held ownership structure continues to be a very common form of organization in the
richest common law countries including the United States. However, Shleifer and Vishny (1986)
11
and Morck et al. (1988) show that the largest American firms have some concentration of
ownership. Shleifer and Vishny (1986) show that such concentration of shareholding can make
sense in terms of solving the agency problem. They argue that if there are many small
shareholders, each will try to take a free-ride on the issue of monitoring the managers. In that
case no monitoring would occur. Hence, if there are large shareholders they would solve the
free-rider problem. In addition, Morck et al. (1988) find that large firms, outside the United
States and the United Kingdom, normally have controlling owners, such as families. These
controlling families maintain their significant influence through various mechanisms such as
pyramidal control structures, cross-shareholdings, and super voting rights. Such mechanisms
allow the families to remain in control even without making commensurate capital investment.
Therefore, given the enormity of these corporations, such families have considerable power in
controlling significant proportions of their countries’ economies.
In a later study, La Porta et al. (1998, 1999) indicate that firms in other developed and
developing countries have a higher level of ownership concentration. The study shows that,
managerial ownership aligns managers’ and outside shareholders’ interests at low levels of
managerial ownership. Up to a certain level of managerial ownership, managers would like to
maximize the firm’s value. However, if the managerial ownership achieves and passes an
optimal level, managers focus on maximizing their own benefits, such as undertaking high-risk
projects, resisting a takeover, and building empires at the expense of other shareholders in the
firm.
The first study to examine the issue of ultimate control is that of La Porta et al. (1999).
Studying ownership structures of large firms in 27 wealthy economies, they find that if we trace
the ultimate owners, we will find relatively fewer firms with widely-held ownership. Even the
largest firms have controlling shareholders, such as families or states. This study underscores the
importance of ownership pyramids through which an ultimate owner could control other
companies by means of indirect ownership. An ultimate owner, who has the most voting rights
(instead of cash flow rights), can be found by tracing the chain of ownership. In addition, the
study shows that different kinds of ownership and control have different impacts on the wealth of
large shareholders.
The idea of ultimate ownership, which was initially propagated by La Porta et al (1999),
instigated a number of related empirical works. First, Claessens, Djankov, and Lang (2000)
12
analyze the separation of ownership and control by using data from nine East Asian countries.
They find that pyramid structures and cross-holdings improve corporate control in all studied
countries. Second, Claessens, Djankov, and Lang (2000) use 1,301 publicly traded corporations
in eight East Asian economies to disentangle the incentive and entrenchment effects of large
ownership. They find that a positive relationship exists between the cash-flow ownership of the
largest shareholder and firm value. However, when the control rights of the largest shareholder
exceed its cash-flow ownership, firm value falls. They also find that managers at family-
controlled firms have more ways to divert benefits to themselves than managers at firms with
widely-held ownership. Third, Faccio and Lang (2002) study the ultimate ownership and control
of corporations. They use data of 5,232 firms in 13 Western European countries. They find that
there are more family-controlled firms than widely-held firms. Also, most financial and large
firms are widely-held firms, while most non-financial and small firms are family-controlled
firms. Furthermore, studies show that in the US and the UK, firms are mostly characterized by
dispersed ownership. However, most of continental Europe and Asia are commonly
characterized by ownership controlled by individuals, families, governments or industrial groups
(La Porta et al 1999, Faccio and Lang 2002). Last, unlike other countries, in China, the
government controls the majority of listed firms (Kato and Long, 2006).
In terms of the nature of governance structures in Canada, Roe and Lee-Sing (1996)
observe that ownership concentration in Canada is high because individuals, families or private
holding companies are the ultimate controlling owners of many large firms. La Porta et al. (1999)
suggest that the ownership structures at Canadian firms are closely similar to those observed in
most countries. Amoako-Adu and Smith (2001) note that the existence of dual-class shares,
which is a relatively common phenomenon in Canada, facilitates concentrated ownership and
family control. Klein, Shapiro, and Young (2005) examine the Canadian evidence on the
relationships between corporate governance, family ownership, and firm value. They suggest
determining the ultimate control by using voting rights instead of equity ownership. They do not
study CEO compensation directly but they test the relationship between firm value and the newly
released indices of effective corporate governance.
2.4.1 Widely-held group versus the concentrated group
13
In our study, we examine the compensation structure of three distinct groups: family-
controlled firms, institution-controlled firms and widely-held firms. We distinguish these three
kinds of firms to represent different levels of market imperfection. Compared with family-
controlled and institution-controlled firms, widely held firms have dispersed ownership. Demsetz
(1983) mentions that since ownership is widely dispersed across many shareholders, no
shareholders can exercise real power to oversee managerial performance in modern corporations.
Shareholders, owning a low amount of shares, have little or no incentives to exert monitoring
behavior (Grossman and Hart, 1988). Thus, managers in widely-held firms have more freedom in
using firm’s capital than managers in non-widely-held firms. Shleifer and Vishny (1997) suggest
that in absence of monitoring, managers would like to maximize their own utilities instead of
shareholders. Another empirical work of Healy and Cole (2002) shows that absence of a
stockholder with a large proportion of stock increases the agency costs, leading to the use of
compensation contracts based on performance.
The most significant weakness of a widely-held ownership structure is the lack of
shareholder monitoring due to the unmatched benefit and cost of monitoring for small
shareholders. The existence of at least one large shareholder will reduce the costs of monitoring
and agency costs and asymmetric information. Major shareholders mitigate the conflict between
managers and shareholders (Shleifer and Vishny, 1986). Shleifer and Vishny (1997) empirically
examine the consequences of corporate ownership for corporate valuation using data on
companies from 27 wealthy countries around the world. They find that compared to small
shareholders, large shareholders have greater resources and incentives to monitor managers
reducing some agency costs.
In particular, large shareholders may be willing to spend time and effort to collect more
information on management performance or to estimate the firm’s investment projects and thus
reduce the information asymmetry. Theoretically, shareholders with significant shares have more
incentives to monitor management. As a result, they (large shareholders) are able to monitor
more efficiently (La Porta et al, 1999). Bebchuk and Stole (1993) suggest if an investor holds a
larger block of shares, this investor will have stronger incentives to protect the investment by
monitoring management. Firth, Fung and Rui (2006) investigate the relationship among agency
costs, ownership structure, and governance mechanisms by using data from China listed firms.
They find that the level of agency cost is not significantly related to individual shareholding,
14
institutional shareholding, and government ownership. Their results indicate that agency costs
are lower because of concentrated ownership, but they are not lower for the boards with a
majority of outside directors. Hence, they argue that though Chinese public firms are undergoing
ownership and governance reforms, such reforms have not yet led to lower agency costs. Their
results support some prior empirical results for the US firms (Singh and Davidson, 2003). A
related study of of Florackis and Ozkan (2008) shows that managerial ownership plays a
significant role for corporate governance mechanism for the UK firms over period from 1999 to
2003. Their results indicate that both compensation and ownership concentration are important
factors in mitigating agency problems. Besides, they find that executive ownership has a positive
relationship with growth opportunities, which means, more growth opportunities firms offer
more incentive mechanisms. Similarly, Ozkan (2007) empirically examines the impact using a
sample of 414 large UK companies for the fiscal year 2003/2004. They also find that institutional,
block-holder ownership, and directors’ are negatively related to CEO compensation
In addition, enhanced monitoring will decrease the waste of free cash flow by managers.
In the USA, publicly traded family-controlled firms (which constitute about one third of the total
listed firms) have higher Tobin’s q values and higher return on assets than nonfamily-controlled
firms (Anderson and Reeb, 2003). Ben-Amar and Andre (2006) state that family ownership has a
positive impact on value creation.
2.4.2 Family-controlled Group Versus Institution-controlled Group
Both family-controlled firms and institution-controlled firms have large shareholders.
However, whether or not the control shareholders are playing an active monitoring role is still an
important issue. Empirical research suggests that institutional investors play an important role on
firm strategies, for example, executive/CEO compensation (Smith 1996; David, Kochhar, and
Levitas, 1998). Smith (1996) concludes that when shareholder activism is successful in changing
governance structure, shareholder wealth will increase. We can argue from Smith’s study that
institutional activism does have impact on share price, which again could affect stock-based
CEO compensation. Similarly, David et al (1998) examine whether institutional investors have
an impact on CEO compensation policy or not. Their results show that institutional owners with
only an investment relationship with a firm influence compensation in accordance with
shareholders preferences to “(1) lower its level and (2) increase the proportion of long-term
15
incentives in total compensation.” Moreover, some other studies have examined the effect of
institution ownership on CEO/executive compensation (Hartzell and Starks, 2003). Hartzell and
Starks (2003) use a sample of 1,914 firms from S&P’s ExecuComp between 1992 and1997.
They find that the concentration of institutional investor ownership is positively tied to the
performance sensitivity of managerial compensation and is negatively tied to the level of that
compensation. They also find a positive relationship between institutional investors and
executive compensation. Besides, their results imply that institution-controlled firms tend to use
incentive compensation to mitigate the agency problem between shareholders and managers.
From the point of aligning the interests of managers and shareholders, the family-
controlled group is superior to the institution-controlled group. According to Jensen and
Meckling (1976), family-controlled firms should be characterized by reduced problems of
agency and agency costs. This hypothesis has been tested and confirmed by Chrisman, Chua, and
Litz (2004). They suggest that the overall agency problems in family-controlled firms is less than
that in non-family-controlled firms using 1,141 small privately held US firms. Demsetz and Lehn
(1985) show that family-controlled firms face less agency problems because they (family-
controlled firms) are able to monitor their managers directly. Similarly, Mehran (1995) examines
the relationship between executive compensation structure and ownership using 153 randomly-
selected manufacturing firms. Results show that firms with a larger percentage of their shares
controlled by outside block-holders offer less long term incentive pay, implying that block-
holder monitoring is a substitute to the incentive components of compensation. Another
empirical work of Kole (1997) indicates that the likelihood of any form of explicit compensation
arrangement is reduced by the presence of an agent of founding family on the board.
First, institutions are more flexible in switching their ownership from one firm to another
depending on performance. If the costs of monitoring are high in comparison to the costs of
rebalancing portfolios, institutions will choose to rebalance instead of monitoring. Unlike
individuals or families, institutions invest money of other people. Institutions have the obligation
to know and protect what they invest. They should take proactive actions, so that the
management of investee firms works towards maximizing shareholder value. (David, Kochhar,
and Levitas 1998). Consistent with internal monitoring of management, substantial top
management changes is negatively related to a firm’s stock returns (Warner, Watt, and Wruch,
1988). They examine this relationship using the sample consists of 269 firms listed on the New
16
York and American Stock Exchanges in the period 1963-1978. Results show that the ratio of the
number of top management changes to the number of firms is relatively stable at 0.183 for all
changes. They indicate that this relationship is consequence from monitoring by the board, other
top managers, or shareholders. Huson, Parrino, and Starks (2002) examine CEO turnover at
large public firms listed in the Forbes over a period of 1971-1994. Results show that the
frequency of forced CEO turnover and the frequency of outside succession are increased. They
also indicate that from the beginning to the end of the period, the relationship between the firm
performance and the likelihood of forced CEO turnover remain the same, even though the
internal mechanisms is significantly changed. Thus, the characteristics of internal monitoring
mechanisms and the nature of CEO turnover do not influence the sensitivity of forced turnover to
firm performance. Kaplan and Minton (2006) studies CEO turnover using data from large US
companies spanning a period of 13 years from 1992 to 2005. The authors find that CEO’s tenure
on average is less than 7 years. Compared to previous studies, the results show that the annual
CEO turnover rate has been increasing. Furthermore, the average tenure of CEOs drops to six
years by using data from 1998 to 2005. They analyze the impact of three components of firm
performance (performance relative to industry, industry performance relative to the overall
market, and the performance of the overall stock market) on internal turnover. Results show that
these three factors have stronger impact on internal turnover after 1998. In contrast, a family that
controls a firm does not have this flexibility.
Second, family-controlled firms generally assign influential positions to family members
whose focus is in line with that of the family group. Even though a non family member may be
appointed as the manager, the level of monitoring is significant given the high ownership
concentration by the family. Anderson and Reeb (2003a) state that family firms are managed or
controlled by founding families. About one-third of the S&P 500 firms across a broad range of
industries are characterized by such ownership. They also suggest that family owners have better
knowledge of the firm’s business activities. Such knowledge helps the owners in detecting
manipulations of stock or firm performance, if any. Bennedsen et al. (2007), using data from
Denmark, show that a professional CEO provides much better performance in a family firm than
a CEO who is a member of the family. Similarly, Villalonga and Amit (2004), upon their study
on all Fortune 500 firms over period 1994-2000, argue that family ownership creates value only
when a founder of the firm or non-family members serve as CEOs. Ben-Amar and Andre (2006)
17
conclude from Canadian data that separation of ownership and control does not negatively affect
value creation. They indicate that family ownership in Canada is a positive factor in value
creation. However, the level of monitoring by a family may not necessarily translate into a
reduction of agency costs for the minority shareholders. Indeed, previous studies suggest that
significant family ownership may lead to agency costs of its own. The family may divert
company resources for its own benefit despite the presence of a manager who may or may not be
a family member. Essentially, the family and the manager can collude to spend on perks and
personal benefits at the expense of minority shareholders. Schulze, Lubatkin, and Dino (2003)
suggest that agency benefits gained by family-controlled firms are offset by free-riding and other
agency problems. Chourou (2010) use a panel of Canadian companies ultimately controlled by
families over a period 2001-2004 to examine hypothesis of owner managers expropriating
minority shareholders by receiving excessive compensation. He suggests that excessive
compensation of chief executive officers at some family owned Canadian corporations may
cause expropriation of minority rights.
2.5 Other Factors
Early studies, for example Baumol (1959) and Lewellen and Huntsman (1970), examined
that factors that determine CEO compensation. They find that firm size and firm performance
have an impact on CEO compensation. Ciscel and Carroll (1980) argue that the findings of these
studies are limited by multicollinearity problems. They find that the market for managerial talent,
the external performance of the firm, and the internal technical efficiency of production affect
the level of executive compensation, after correcting for multicollinearity.
Some studies examine the relationship between CEO compensation and market or
industry performance. Hart (1983) indicates that when a manager owns a small stake the product
market may still force managers to follow the principle of shareholder value maximization.
Hart’s argument is based on the assumption that a given product has a cost component that is
common among all producers. If agency costs make the costs of a product higher than the costs
of its peers, consumers will avoid buying it. This result affects negatively the manager’s personal
benefits.
Jensen and Ruback (1983) focus on the role of the market for corporate control. They
investigate the relationship between managers and shareholders and corporate takeovers. Their
18
results show that corporate takeovers benefit the target firm’s shareholders. In addition, the
shareholders in the bidding firm do not lose. More recently, Tannous and Cheng (2007), propose
that the market for corporate control provides another incentive for managers to perform. They
provide evidence suggesting that corporate takeovers are often motivated by the poor
performance of the target and turn around plans that include dismissal of existing managers.
Gibbons and Murphy (1990) argue that corporate performance depends on non-
controllable factors such as industry and market conditions, which also have an impact on CEO
compensation. They examine the relationship between relative performance and CEO
compensation. Their empirical evidence strongly supports the existence of a positive relation
between CEO pay and firm performance but they find that CEO pay is negatively related to
industry and market performance. They also find that CEO performance is more tied to
aggregate market movements than industry movements.
Jacobs (1991) shows that CEO compensation is correlated to market effects. He argues
that managerial short-sightedness is a key reason for the decline in American business
competitiveness. Such short-sighted behaviour of managers raises questions regarding whether
better designed compensation contracts could induce managers to behave in a way that is
consistent with the long-run interests of shareholders.
19
CHAPTER 3
Theoretical Arguments and Hypotheses
Past studies indicate that the ownership structure affects the degree of agency costs. It is
argued that the higher the ownership of a particular entity the lower will be this entity’s
monitoring costs in proportion to its benefits of monitoring the managers. We propose that the
ownership structure is a spectrum that ranges from full ownership by one individual to widely
dispersed ownership by a large number of shareholders each owning a very small portion of the
firm. Furthermore, we propose that the costs of monitoring management effectively are
significant but these costs are fixed while the benefits of monitoring management are
proportional to the percentage of ownership in the firm. Thus, shareholders who have significant
ownership in a firm should be willing to monitor managers closely, which suggests that
concentrated ownership in a firm should increase the level of monitoring and reduce agency
costs.
Similarly, previous studies suggest that the compensation contract can be designed to tie
a portion of the pay to performance. It is argued that performance-based compensation will align
the interests of managers and shareholders. Therefore, it should reduce the need for shareholders
to monitor the performance of managers. Therefore, if we assume that shareholders are indeed in
a position to structure compensation packages as they please regardless of the ownership
structure then we can suggest the existence of a negative relation between the degree of
monitoring and the proportion of performance-based CEO compensation. Under this scenario the
performance-based compensation may vary depending on the ownership structure of firms. The
widely-held firms, in which ownership is widely dispersed across many shareholders, should rely
heavily on performance-based compensation to align the interests of managers and shareholders
and reduce agency costs. In contrast, firms which have concentrated ownership should have less
need for performance-based compensation. In these firms, the existence of at least one
shareholder with a significant ownership stake will improve monitoring and reduce the level of
agency costs.
However, there is no evidence that suggests shareholders are in control of compensation
packages. In the contrary, the evidence suggests that executive compensation is mainly
determined by competitive pressures in the market for CEOs and by benchmarking. We propose
that in this environment concentrated control by an institution or a family may or may not lead to
20
significant differences in incentive compensation across ownership structures. However, in the
presence of high monitoring by families or by institutions we should find a significant relation
between incentive compensation and performance.
Previous studies suggest that performance based compensation can take many shapes and
forms. The annual compensation of a CEO in a typical firm has three components that account
for approximately 90% or more of the CEO’s total compensation.1
The three most significant
components are fixed annual salary, bonus, and contingent compensation (contingent
compensation consists of stock options, performance plans, restricted options, and other long-
term incentives). The bonus is usually based on some accounting metric such as return on assets,
return on equity, or cash flow per share. Theoretically, these metrics are positively related to the
value of the firm and good performance along these metrics will improve shareholder’s value.
Furthermore, it is easy to link performance in accounting measures to managerial actions.
Therefore, it is preferable in environments of active monitoring. However, the bonus as an
incentive pay may be criticized on the basis that it focuses the attention of managers on short-
term performance and distracts from capitalizing on the long-term interests of shareholders.
Stock-based compensation is introduced to align the long-term interests of shareholders with the
interests of managers. It is argued that managerial equity ownership in a firm leads CEOs to
manage in the best interests of shareholders because these interests are their own interests.
However, there are wide differences of opinion among academics, practitioners, and policy
makers regarding the effectiveness of stock-based compensation in promoting the long-term
interests of shareholders. First, stock-based compensation may reward managers for success
resulting from factors beyond the control of managers. Second, managers may not hold enough
cash flow interest to align their interests with those of the owners. Third, there is no universally
accepted model or method that can help owners to set the optimal stock-based compensation
component to prevent excessive compensation. Therefore, we propose that shareholders who are
in control of compensation packages are likely to prefer bonus compensation as a way to control
the actions of managers and they are likely to offer equity based compensation based on market
practices benchmarking methods.
1 The remaining portion consists of annual pension, annual vocation, lump-sum pension, and retirement
allowances.
21
In summary, our hypotheses are as follows:
H1a: The proportion of CEO annual bonuses in concentrated firms is lower than the
proportion of CEO annual bonuses in widely-held firms.
H1b: The proportion of CEO contingent compensation in concentrated firms is lower than the
proportion of CEO contingent compensation in widely-held firms.
Concentrated ownership may be the result of high percentage ownership by a family or
by an institution. We propose that either form of control should improve monitoring and reduce
agency costs but if all else are equal control by a family provides better monitoring than control
by an institution. Therefore, our hypotheses can be expanded to include:
H2a: The proportion of CEO annual bonuses in family-controlled firms is the largest among
the three types of ownerships: family-controlled, institution-controlled, and widely-held.
H2b: The proportion of CEO annual bonuses in institution-controlled firms is lower than that
in family-controlled firms and is higher than that in widely-held firms.
H2c: The proportion of CEO annual bonuses in widely-held firms is the lowest among the
three types of ownerships: family-controlled, institution-controlled, and widely-held.
H3a: The proportion of CEO contingent compensation in family-controlled firms is the lowest
among the three types of ownerships: family-controlled, institution-controlled, and
widely-held.
H3b: The proportion of CEO contingent compensation in institution-controlled firms is higher
than that in family-controlled firms and is lower than that in widely-held firms.
H3c: The proportion of CEO contingent compensation in widely-held firms is the highest
among the three types of ownerships: family-controlled, institution-controlled, and
widely-held.
Previous empirical studies on CEO compensation concentrate mainly on the pay to
performance relationship. Earlier studies, for example Jensen and Murphy (1990), Kaplan (1994),
and Hall and Liebman (1998), argue that incentive pay is related to firm performance. However,
22
recent studies suggest that the relation between performance and incentive pay is weak. Shaw
and Zhang (2010) find that CEO bonus compensation is less sensitive to poor earnings
performance than it is to good earnings performance. Similarly, Fahlenbrach and Stulz (2011)
find no evidence to support the proposition that the performance of banks during the 2008
financial market crisis is positively related to the alignment of incentives of bank managers and
their shareholders. We examine the relation between incentive pay and firm performance to see
whether controlling for ownership structure can clarify this relation.
H4a: Annual bonus of CEOs is positively related to firm performance and this relation varies
across ownership structures
H4b: Contingent compensation of CEOs is positively related to firm performance and this
relation varies across ownership structures
We examine how incoming CEOs are compensated in comparison to their predecessors.
We propose that experience in the industry and in similar position may enable the incoming CEO
to negotiate a high compensation package and a structure that is in the best interests of the CEO.
On the other hand, the departure of a CEO could be seen as an opportunity for a firm to re-
establish its own priorities and to design the compensation package to promote the interests of
shareholders and the ultimate power brokers of the firm. Accordingly, we expect that the
compensation of the incoming CEOs would be structured differently than the compensation of
their predecessors. Furthermore, we propose that the structure of the compensation packages of
incoming CEOs will vary depending on the ownership structure.
H5a: The bonus compensation of incoming CEOs is lower than the bonus compensation of
their predecessors. H5b: The contingent compensation of incoming CEOs is higher than the contingent
compensation of their predecessors.
23
CHAPTER 4
Data
4.1 Ownership Structure
The sample selection starts by considering the 269 firms that made up the S&P/TSX
Composite Index (formerly known as the TSX 300) at the end of 2007. These firms comprise
approximately 71% of the market capitalization of all Canadian-based companies listed on the
TSX. Our data spans the years 2003 to 2007 inclusive. Because the components of the S&P/TSX
index vary from year to year, we choose the components of the 2007 index as the initial sample
and then we track back through the sample period to select only the firms that have continuous
presence on the S&P/TSX index throughout the sample period. By doing this, we get 143 firms
which have full records.2
Data on ownership structure are manually collected from the Inter-Corporate Ownership
(ICO) database which is released quarterly by Statistics Canada.
3
From the ICO, we find that there are three overwhelming kinds of ownership structures:
family-controlled, institution-controlled, and widely-held. The widely-held group consists of all
companies that do not have a controlling interest of 10% or more. We define a firm as family-
controlled if its controlling shareholders who own 10% or more are from a family-controlled
group. Institution-controlled are firms with controlling shareholders from institutions such as
pension plans, mutual funds, trusts, banks, and insurance companies. The categorization between
widely-held, institution-controlled, or family-controlled is done each year from 2003 to 2007
inclusive which allows ownership to change over time. We observe that compared to the other
two groups, the family-controlled firms have the most stable ownership structure over time.
We choose the last quarter of
each year to represent the firms’ ownership for the entire year. We rely on the annual reports
instead of the quarterly reports because the quarterly reports contain missing data. The
accounting data is collected from the Compustat database.
In our study, a firm is categorized as having controlling shareholders if 10% of the voting
shares of the firm are owned by a single individual, a group of individuals acting together, a
2 Our data sample may be criticized on the basis that the method by which we selected the firms will subject the
results to survivorship bias. However, our main objective is to compare the incentive compensation among the three ownership groups. We expect that the survivorship bias, if any, will have the same impact on each of the three groups. Thus, the survivorship bias will not affect the comparison results.
3 Statistics Canada requires publicly held companies to identify any controlling interest of 10% or more. Failure to comply with this requirement violates Canadian laws and subjects the violators to penalties.
24
family, an institution, or another corporation. Previous studies, for example La Porta et al. (1999),
show that this threshold is sufficient to control a firm’s decision making system. Furthermore,
other Canadian databases, such as the Financial Post (FP) Informat, use this threshold to classify
firms between concentrated ownership and dispersed ownership.
While the 10% cut off is sufficient, it may not be necessary. Critics may argue that the 10%
cut off is too high as ownership interests less than 10% may be effective in controlling a
company. We agree with this argument but we cannot lower the cut off ownership level. The
ICO database from which we obtain the control information does not report on ownership
interests less than 10%. We decided to rely on the ICO despite its limitations for two reasons.
First, we feel that it is the most reliable source given that the information is collected to comply
with government regulations. Second, we considered the possible effects of using a high cut off
level and concluded that it is not affecting the qualitative results. In our view, using a 10% cut
off level instead of a lower one, for example 5%, would improperly classify some firms as
widely-held when they should be classified as family-controlled or institution-controlled. The
impact of this misclassification would be to weaken the differences between the widely-held
firms and each of the other two groups. Our results suggest that the differences are significant
despite the possible misclassifications.
According to La Porta et al (1999), an ultimate owner is the entity that has the most
voting rights instead of cash flow rights. For example, a family-controlled firm is controlled by
an individual or family owning 10% or more of the firm’s voting stock. For this purpose, we
track the ownership structures of the direct controllers and categorize firms accordingly. For
example, Ensign Resource Service Group Inc. is controlled by the Mackenzie Financial
Corporation through 12.33% of the voting rights. Meanwhile, Mackenzie Financial Corporation
is 100% controlled by a family-controlled group. Under this case, Ensign Resource Service
Group Inc. is categorized as family-controlled firm. In our sample, there are several companies
for which the direct controllers are institutions while the ultimate controllers are families.
4.2 CEO compensation
CEO’s compensation data are hand-collected from the proxy circulars of each company
as listed in the System of Electronic Document Analysis and Retrieval (SEDAR)4
4 SEDAR is a comprehensive, online archive of securities documents filed by publicly traded companies in Canada.
. During our
25
study period, the majority of sample firms (87) had the same Chief Executive Officer over the
entire 5-year period. We call this subsample the Permanent CEO Group. The remaining 39 firms
experienced one or more changes in CEOs during the study period. We call this subsample the
Transient CEO Group. Since 8 firms changed ownership structures in the transient CEO group,
we use 31 (39 minus 8) firms to compare the compensation of the new CEOs with the
compensation of their predecessors.5
We combine the Permanent CEO Group and the Transient CEO Group to conduct
multivariate analysis. For the 87 firms that make up the Permanent CEO Group, we obtain the
market-to-book (M/B) ratio, the return on assets (ROA), and the debt-to-equity (D/E) ratio of
each company from the Compustat database. For 8 firms, ratios are not reported. Thus, for the
subsample of permanent CEO firms, we have 79 firms. In the Transient CEO Group, ratios of 6
firms are not available. Therefore, for the subsample of transient CEO firms, we have 33 firms.
As the study covers the period 2003-2007 inclusive, for this group we have a total of 560
observations. The composition of the observations is as follows: 96 from family-controlled firms,
228 from institution-controlled firms, and 236 from widely-held firms.
Eight firms changed CEOs more than once during the 5-
year period of our study.
The components of CEO compensation vary among various companies. In this study, the
analysis of CEO compensation is based on five different measures of compensation: total
compensation, salary, annual bonus, contingent compensation, and all other compensation.
Salary measures the component of compensation that is fixed at the beginning of the year. The
annual bonus is the short-term incentive which is often based on accounting measures of
performance. Contingent compensation consists of stock options, performance plans, restricted
options, and other long-term incentives6
5 The compensation of departing and the incoming CEOs are reported separately in the compensation information
we found in the proxy circulars. If firms report compensation for a portion of the year for the departing and the incoming CEOs, we annualize the compensation.
. The term ‘contingent’ indicates that such compensation
depends on the performance of the underlying asset. All other compensation consists of annual
6 Some firms specify the maximum, target, and threshold of future payouts A firm grants a number of shares to its CEO, and it sets performance goals for the CEO. If firm performance is superior, the CEO can get a contain percentage (e.g.120%) of grant shares, which is the Maximum. If firm performance is average, the CEO gets the same number of shares as they are granted, which is the target. If firm performance is below the required level, the CEO will only get a percentage less than the target (e.g. 80%), which is threshold. We use the target to estimate the compensation.
26
pension, annual vocation, lump-sum pension, and retirement allowances. Total compensation is
the sum of salary, annual bonus, contingent compensation, and all other compensation.
As salary, annual bonus and all other compensation are typically paid in cash or cash
equivalent, the valuation of these three components is straightforward. However, the valuation of
the contingent component is complex. We use the Black-Scholes Option Pricing model to
calculate the dollar value of total contingent compensation.7
Furthermore, to be consistent with prior studies (Jensen and Murphy 1990a and Zhou
2000), we use the standard deviation of the continuously compounded monthly return as our
volatility. We use the monthly total return index over a three-year period ending with the grant
date. We obtain this data from DataStream. This approach may be criticized on the basis that
weekly and annual return may be more appropriate. First, weekly returns have higher
autocorrelations than monthly and annual returns. Second, using annual returns to calculate
volatility requires data from prior years and that may reduce the number of companies in our
sample due to missing data. Third, the volatility as determined from prior years, many of which
may be far from the grant date, may not reflect the true volatility of the underlying security.
This process involves several steps.
First, we need to know the number of options granted the exercise price, and the time to expiry.
These data are collected from the proxy statements. Second, we use the 3-month Treasury bill
rate as the risk-free rate. We obtain this data from the Bank of Canada. Third, we assume the
time period is monthly. Then, if companies provide the grant date only for the grant year, but do
not provide a grant date in later years, we assume that the grant date in later years is the same
date as the grant year. Fourth, if companies do not give the grant exercise price, we assume that
the closing share price on the grant date is the grant exercise price.
Moreover, a number of firms report CEO compensation annually in US dollars. Some of
these firms also report the average exchange rate for the year. For these firms, we convert the
CEO compensation from US dollars to Canadian dollars using the reported exchange rate. If
firms reported CEO compensation annually in US dollars, but do not report the exchange rate,
we use the average exchange rate for the year as reported by the Bank of Canada. In addition,
7 Canadian firms under the TSX do not need to report the values of option grants. We estimate the monetary
values of options by the Black-Scholes formula (Fischer Black and Myron Scholes 1973). Option value 𝐶0 = 𝑆0 × 𝑁(𝑑1) − 𝐸/(1 + 𝑅𝑓)𝑡 × 𝑁(𝑑2), where 𝑑1 = �ln(𝑆0 𝐸⁄ ) + �𝑅𝑓 + 1/2 × 𝜎2� × 𝑡�/�𝜎 × √𝑡�, and 𝑑2 = 𝑑1 − 𝜎 × √𝑡. In the formula, 𝑆0 is the value of the stock at the date, 𝐸 is exercise price, 𝑡 is expiration term (in months), 𝑅𝑓 is the risk-free interest rate on 3-month Canadian Treasury Bills, 𝜎 is volatility, and N(•) is the cumulative standard normal distribution function.
27
with the intention of comparing compensation of departing and incoming CEOs, we record
salary, bonus, and contingent pay in the year of the grant for both departing and incoming CEOs
in the Transient group.
4.3 Variables
4.3.1. Dependent variables
We analyze three measures of CEO compensation: annual bonus, contingent pay, and
total pay. Contingent pay includes securities under options (SUO), stock appreciation rights
(SAR), value of restricted share units (LTIP), and contingent incentive pay. Total pay includes
salary, annual bonus, contingent pay, and other pay. In addition, we use the percentage of each
category of compensation (out of total compensation) as our dependent variable. In addition, we
use the compensation per dollar of total assets as our dependent variable.
4.3.2. Control variables
Our review of prior research suggests the inclusion of four control variables in our
analysis. Total assets represent one measure of size. This measure was used by Daily et al. (1998)
and Chowdhury and Wang (2009). Chowdhury and Wang (2009) found a positive relationship
between total assets and executive compensation in Canadian corporations. Therefore, in our
original model, we control for firm size by using the natural logarithm of assets.
We use Tobin’s q ratio to control for the growth opportunities of the firm. Tobin’s q
measures the market value of a firm’s assets in relation to their replacement cost. Many studies,
for example Hartzell and Starks (2003) and Harvey and Shrieves (2001), document a strong
relationship between growth opportunities and the presence of incentive compensation. In
addition, Chowdhury and Wang (2009) show that growth opportunities are positively related to
incentive compensation of institution-controlled firms in Canada. We use Tobin’s q to account
for the variation in firm performance and to control for the presence of growth opportunities. We
use the market-to-book (M/B) ratio as a proxy for Tobin’s q. We download the appropriate data
from the Compustat database.
Consistent with David et al (1998) and Chowdhury and Wang (2009), we use the return
on assets (ROA) to measure firm performance. Chowdhury and Wang (2009) did not find a
relationship between financial performance and executive compensation of Canadian firms while
28
Zhou (2000) suggests that an overall weak relationship exists between executive pay and
performance in Canadian firms.
We use the debt to equity ratio as one of the explanatory variables. Previous studies, for
example Healy and Cole (2000), argue that CEOs of firms with a high level of indebtedness
would prefer less contingent pay to avoid the increased risk, because the amount of cash
available for either dividends or cash compensation is influenced by debt. However, owners
would like to offer CEO more stock-based compensation to prevent the CEO from choosing a
debt to total equity ratio that is suboptimal to the stockholders. Hence, we expect a positive
relation between the debt to equity ratio and incentive pay.
4.3.3. Dummy variables
Our main concern objective is to examine the relation between the ownership structure
and CEO compensation. We start by comparing CEO compensation at the concentrated group
with CEO compensation at the widely-held group. Dcon is the dummy variable that takes the
value of 1 if a firm belongs to the concentrated ownership group and 0 otherwise. Then, the
sample is divided between widely-held, institution-controlled, and family-controlled with the
widely-held group serving as the base group. DF is the dummy variable that takes the value of 1
if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value
of 1 if a firm is institution-controlled and 0 otherwise. After that, we switch the base group from
widely-held group to institution-controlled group, while DW is a dummy variable that takes the
value of 1 if a firm is widely-held group and 0 otherwise.
In addition, we examine the industry effects. Given the limited sample size, we divide our
sample into four industries based on the list of companies in the TSX Sector Indices: S&P/TSE
Canadian Energy Sector Index, S&P/TSE Canadian Financials Sector Index, S&P/TSE Canadian
Materials Sector Index, and other Indices.8
8 The other S&P/TSX sector indices are: Canadian Consumer Discretionary Canadian Consumer Staples, Canadian Diversified Metals & Mining, Canadian Gold Index, Canadian Health Care, Canadian Industrials, Canadian Information Technology, Canadian Real Estate, Canadian Telecommunication Services Sector, and Canadian Utilities
We use other indices as the base index and add three
dummies to represent the financial industry (DFin), the energy industry (DEgy), and the material
industry (DMat). DFin is a dummy variable that takes the value of 1 if a firm is a member of the
financial industry and 0 otherwise. DEgy and DMat are defined in the same manner.
29
We use another dummy variable, D1YC, to control for newly appointed CEOs. D1YC
takes the value of 1 if the CEO is in her/his first year on the job and 0 otherwise.
Finally, we control for the year effects. We use year 2003 as the base year and add the
four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005,
2006, and 2007.
30
CHAPTER 5 Descriptive Statistics and Univariate Tests
In this chapter, we examine the trends of CEO monetary compensation over the period
2003-2007. For this analysis, the focus is on the compensation of CEOs who remained in the
same position throughout the study period. Each component of CEO compensation is calculated
as a percentage of total pay. In addition, for transient CEO firms we compare CEO compensation
before and after the turnover. The analysis is conducted for the three different ownership
structures: family-controlled, institution-controlled, and widely-held. Finally, for both permanent
CEO firms and transient CEO firms we present descriptive statistics and univariate tests to
compare the family-controlled, institution-controlled, and widely-held firms.
5.1. Permanent CEO firms
The following tables and figures analyze CEO compensation across different years. Table
1.1 presents the salary, annual bonus, contingent pay, and total compensation paid by permanent
CEO firms. The table shows that all components of compensation have been increasing in
Canada over the five-year period of 2003-07. We find that the family-controlled group has the
highest pay in terms of salary, annual bonus, contingent pay, and total compensation. In addition,
the growth rate of the CEO compensation is higher in the family-controlled group than in the
other two groups. In 2003, the institution-controlled group has the lowest pay among the three
groups. All components of CEO compensation have been increasing steadily from 2003 to 2006.
This finding is consistent with the conclusions of previous studies which show that the
contingent component of CEO compensation has increased in Canada (Zhou 2000; Chowdhury
and Wang 2009). From 2006 and onwards, the components of compensation in the institution-
controlled group were higher than the components of compensation in the widely-held group. In
widely-held group, we see that the salary and annual bonus have increased sharply, but there is
no change in the contingent pay. Thus, in 2007, the widely-held group has the lowest pay among
the three groups.
[Insert Table 1.1 Here]
These findings can be observed in Figures 1.1, 1.2, and 1.3. Figure 1.1 shows that salary9
9 We obtain the annual rate of inflation from the Bank of Canada website.
,
adjusted for inflation, has been continually increasing from 2003 to 2007 for every ownership
31
group. Compared to the institution group and the widely-held group, the family-controlled group
offers the highest salary for the CEO. Figure 1.2 shows that the annual bonuses of the widely-
held and institution-controlled firms have moved higher slowly and are almost equal in 2007. In
contrast, the annual bonuses of the family-controlled group have increased at a faster rate
steadily over the years. Figure 1.3 shows that contingent compensation of the widely-held firms
is almost stable during our study period. In contrast, the contingent pay of the family firms has
increased sharply between 2003 and 2005 and stabilized in 2006 and 2007. Similarly, contingent
pay of the institution-controlled group increased steadily between 2003 and 2006 but declined in
2007. Figure 1.4 illustrates that the total pay of both the family-controlled and the institution-
controlled groups has an upward trend, whereas total pay of the widely-held firms is almost
constant.
[Insert Figures 1.1, 1.2, 1.3, and 1.4 about Here]
Table 1.2 presents each of the components of CEO compensation as a proportion of total
pay. The table shows that salary makes up a small proportion of total compensation. In 2003, the
proportion of salary to total pay was 16.53%, 18.83%, and 16.45% respectively for the family-
controlled, institution-controlled, and widely-held firms. Between 2004 and 2007 inclusive the
ratio declines for the first two groups and remains almost stable for the widely-held firms.
[Insert Table 1.2 Here]
The table also shows that the ratio of CEO annual bonus to total pay in 2003 was almost
equal to the ratio of salary to total pay. In 2003, bonus compensation was 19.37%, 17.13%, and
18.01% of total pay respectively for the family-controlled, institution-controlled, and widely-
held firms. However, contrary to salary the proportion of bonus pay increased on average after
2003. In 2007, bonus payments account for 31.60%, 20.07%, and 24.46% of total compensation
in the family-controlled, institution-controlled, and widely-held firms. In particular, the
proportion of bonus payments in the family-controlled group experiences a sharp rise from 19.37%
in 2003 to 31.60% in 2007. In the institution-controlled group the ratio in 2003 was 18.01% and
it increased by only 2.8% over the five-year period. In the widely-held group, bonus
compensation moves up to 24.46% over the five-year period. Hence, we can conclude that the
annual bonus payments as proportion of total compensation have increased from 2003 to 2007
regardless of the ownership structure to which a firm belongs.
32
Furthermore, Table 1.2 shows that contingent pay makes up a significant proportion of
total CEO compensation. In 2003, contingent compensation accounted for 58.55%, 57.81%, and
63.95% of total pay respectively for the family-controlled, institution-controlled, and widely-
held firms. These proportions fluctuate slightly between 2003 and 2007 but contingent
compensation remained well above 50% of total pay throughout the 5-year period. For example,
the proportion of CEO contingent compensation in the family-controlled group increases from
58.55% in 2003 to 60.19% in 2005, but thereafter decreases to around 53%. Similarly, the ratio
for the institution-controlled group rises to 68.15% in 2006, but it decreases to 60.94% in 2007.
In the widely-held group, the proportion of contingent compensation in total pay decreases
steadily from 63.95% in 2003 to 52.38% in 2006 but rises to 55.11% in 2007.
Figures 2.1, 2.2, and 2.3 present each of the components of CEO compensation as a
proportion of total pay. Figure 2.1 illustrates that although S-to-TP in the institution-controlled
group is the highest in 2003, it decreased continually from 2003 to 2006. The S-to-TP has
become lower than the ratio in the widely-held group since 2006. S-to-TP in the widely-held
group is relatively stable. In the family-controlled group, the S-to-TP deceases sharply in 2003.
However, the ratio remains unchanged from 2004 and onwards.
Figure 2.2 shows the annual bonus as a proportion of total compensation. We see that the
family-controlled group offers the highest proportion for the CEO across the three groups. Also,
we see an upward trend of AB-to-TP over the five-year period. In widely-held group, AB-to-TP
moves downward from 2006. In the institution-controlled group, we observe that AB-to-TP rises
sharply from 2003 to 2004, decreases the next two years and finally recovers back in 2006.
Figure 2.3 shows that CP-to-TP follows a similar trend in the family-controlled group and
the widely-held group over the study period. In particular, CP-to-TP follows a constant pattern
from 2003 to 2005, a sharp decline for some time period and remains unchanged afterwards. In
the institution-controlled group, CP-to-TP has increased dramatically from 2005 to 2006, and
then starts to fall in 2006.
[Insert Figures 2.1, 2.2, and 2.3 about Here]
5.2 Transient CEO Group
A total of 39 firms (31% of all consistent ownership structure firms) replace their CEOs
within the 5-year period. This result implies a replacement rate of 6% of CEOs per year. This
33
turnover is consistent with other studies that look at CEO turnover around this time frame by
using US data (for example, Kaplan and Minton, 2006; Chhaochharia and Grinstein, 2009).
Since 8 firms change both ownership structures and CEOs, 31 transient CEO firms are used to
compare the compensation packages of new CEOs with the packages of their predecessors. In
the multivariate tests, we include these 8 firms in the Transient CEO Group.
Regarding the corporations that change CEOs, we treat the change event for every CEO
as a way to divide the data into two sub-samples. In the case of two changes, we have three
periods. One period before the first change, another period after the second change, and the third
period is in the middle. We give each company a maximum of one change. In the case of two or
more CEO changes, the compensation of the first departing CEO is added to the data before the
change and the compensation of the last incoming CEO is added to the data after the change.
That is, we compare the last period with the first and ignore the periods in the middle.
We find that in the family-controlled group, firms offer the new CEO contingent
incentive plans at the beginning of the term. As a result, contingent compensation and total
compensation of the first year are much higher than other years.
Figure 3.1 presents the annual salary, annual bonus, contingent pay, other compensation,
and total compensation of new and departing CEOs of the family-controlled firms. Total
compensation of the departing CEOs includes payments related to retirement. The figure shows
that the new CEOs in the family-controlled group get, on average, higher compensation in terms
of contingent compensation and total compensation. This observation is consistent with prior
studies. Chowdhury and Wang (2009) suggest that the fortunes of many Canadian companies
have climbed during the past few years because of the increasing price of oil. This could be one
of the reasons why compensation of the new CEOs is higher than the compensation of the old
CEOs.
[Insert Figure 3.1 about Here]
Figure 3.2 illustrates the annual salary, annual bonus, contingent pay, other compensation,
and total compensation of new and departing CEOs of the institution-controlled firms. It shows
that in the institution-controlled firms, the new CEOs receive almost the same salary but higher
bonus and higher contingent compensation than their outgoing peers. As a result, total
compensation is higher as well.
[Insert Figure 3.2 about Here]
34
Figure 3.3 depicts the annual salary, annual bonus, contingent pay, other compensation,
and total compensation of new and departing CEOs of the widely-held firms. It shows that in
these firms, the new CEOs receive almost the same salary but lower bonus and lower contingent
compensation than their outgoing peers. As a result, total compensation is lower as well.
[Insert Figure 3.3 about Here]
Our sample contains five firms whose CEOs retired during our study period. Specifically,
four of these CEOs retire while employed at institution-controlled firms and one CEO retires
while employed at a family-controlled firm. Figure 4.1 reports the same information as in Figure
3.1 after deleting the family-controlled firm whose CEO was replaced due to retirement. The
table shows that the result shown in 3.1 remains unchanged. One exception is that the new
CEOs get more other pay than old CEOs. Figure 4.2 reports the same information as in Figure
3.2 after deleting the four institution-controlled firms who’s CEOs were replaced due to
retirement. The figure shows that, on average, incoming CEOs receive higher annual bonus,
contingent pay, and total compensation than their predecessors. At the same time, incoming
CEOs are paid similar salary as the outgoing CEOs while other compensation of the incoming
CEOs is less than other compensation of the outgoing CEOs.
[Insert Figures 4.1 and 4.2 Here]
Table 2.1 presents descriptive statistics for all components of compensation (annual
salaries, annual bonuses, contingent compensation, and total pay) reported for the departing
CEOs and for the incoming CEOs. On average, the salary of departing CEOs is higher than the
salary of incoming CEOs. However, all other components of compensation are higher for the
incoming CEOs than those for the departing CEOs. Total compensation of the incoming CEOs is
also higher than the total compensation of the departing CEOs.
[Insert Table 2.1 Here]
Panel A of Table 2.2 shows the t-test results of annual bonus and contingent
compensation in the Transient CEO Group. The evidence shows that the compensation, in terms
of bonus, contingent pay, and total pay, of incoming CEOs is not significantly different from the
compensation of outgoing CEOs.
[Insert Table 2.2 Here]
The table also presents the results of comparing the salary, bonus, and contingent pay as
percentages of total pay of incoming and departing CEOs. We find that the contingent pay as a
35
percentage of total compensation of incoming CEOs is significantly higher than contingent pay
as a percentage of total compensation of outgoing CEOs.
In conclusion, we find that the contingent pay of incoming CEOs is higher than the
contingent pay of their predecessors particularly in the family-controlled and the institution-
controlled firms. In contrast, in the widely-held firms contingent pay of new CEOs is lower than
the contingent pay of their predecessors. This finding is consistent with Hypothesis H5b.
5.3 Aggregate Sample
5.3.1 Descriptive Statistics
Table 3 presents the summary statistics of the aggregate sample. We examine both the
dependent and control variables. Our results show that contingent pay has a higher proportion of
total pay than that of annual bonus. Also, some firms do not offer their CEO contingent pay in
some specific years. The mean of the natural logarithm of firm size is 21.77 indicating that firm
size is on average 2.848 billion dollars in assets. We use the variable return on assets (ROA) to
measure the firm’s performance. Some firms report negative ROA which means they do not
perform well during our study period. The mean ROA is 4.17%. Market-to-book ratio (M/B) is
used as a proxy for the firm’s investment opportunities. The average of M/B is 2.73. Firms with
above average M/B indicate more investment opportunities. We expect that the proportion of
CEO contingent pay in such firms would be higher than firms with less investment opportunities.
The debt-to-equity (D/E) ratio can have an impact on the availability of cash. We find that the
mean of the D/E ratio is 85.77%.10
[Insert Table 3 Here]
Table 4 shows the summary statistics of both the dependent and control variables for each
of the three ownership structure groups. Several observations can be made. We find that the role
of different components of CEO pay tends to be different across firm size. As firms become
larger, all components of compensation become higher. First, the family-controlled group has the
largest average firm size and the highest compensation pay. Second, the family-controlled group
has the highest average D/E ratio. Third, the family-controlled group has the lowest ROA and
lowest M/B ratio. Fourth, the institution-controlled and widely-held firms have similar average
firm size, average D/E ratio, average ROA, and average M/B. Fifth, the widely-held firms and
10 Quebecor World Inc has an unusual D/E ratio of 1383.31%. We ran regressions after excluding this observation from the data. The results remain unchanged. Our reported results are based on the sample that includes Quebecor.
36
the institution-controlled firms pay similar compensation in the form of annual bonus, contingent
pay, and total pay.
[Insert Table 4 Here]
We separate our sample into four different industries and Table 5 presents the summary
statistics of both dependent and control variables for each industry. There are 90 observations
that belong to the energy industry, 90 observations that belong to the financial industry, 110
observations that belong to the materials industry, and 270 observations in other industries.
Energy, material, and financial industries are important industries in Canada, constituting 16%,
16% and 20%, respectively, of the total sample. Several observations can be made from the table.
First, firms in the financial industry have the largest average size while the average firm size in
the energy and material industries and other industries have similar average firm sizes. Second,
the executives in the financial services industry earn higher pay than all other industries. In
particular, financial CEOs receive notably higher pay in terms of the annual bonus and
contingent pay. Third, firms in the energy and material industries have relatively lower pay than
firms in other industries. Fourth, the financial industry has the lowest average ROA (2.34) while
the ROA of the energy and material industries is almost three times higher. The ROA in the other
firms is approximately 3.34. Fifth, the material industry has the highest average M/B of 3.06
followed by other industries (2.7) and the energy industry (2.6) while the financial industry has
the lowest M/B of 2.5. Sixth, the material industry has the lowest average D/E ratio while the
financial industry has the highest average D/E ratio.
[Insert Table 5 Here]
5.3.2 Univariate Tests
We conduct t-tests to determine the significance of the differences in the compensation
levels paid by the family-controlled, institution-controlled, and widely-held firms. In particular,
we conduct t-tests to compare bonus, contingent, and total compensation. The results are
reported in Table 6.1. The table shows that both annual bonus and contingent pay of the family-
controlled firms are significantly different from their counterparts at the institution-controlled
firms or the widely-held firms. The family-controlled firms and institution-controlled firms seem
to be using different levels of incentive payments to motivate their CEOs although both have
high degree of concentration of ownership. Furthermore, Table 6.1 shows that the differences in
37
the components of compensation between the institution-controlled group and the widely-held
group are not statistically significant.
[Insert Table 6.1 Here]
The results reported in Table 6.1 may be biased by size. We conduct z-tests to determine
the significance of the differences in the structure of compensation paid by the family-controlled,
institution-controlled, and widely-held firms. In particular, we conduct z-tests to compare bonus
and contingent pay as percentages of total compensation. The results are reported in Table 6.2.
The table shows that the family-controlled firms pay the highest bonus per dollar of
compensation ( 23.6%) while the institution-controlled firms pay the lowest bonus per dollar of
total compensation (21.6%). The widely-held group has a slightly higher percentage (21.9%)
than the institution-controlled group. However, the differences among the ratios are statistically
insignificant.
[Insert Table 6.2 Here]
Table 6.2 also shows that the family-controlled and the institution-controlled firms pay
higher contingent compensation as percentage of total compensation than the widely-held firms.
The institution-controlled firms pay the highest proportion of compensation in the form of
contingent pay (46.8%). The family-controlled firms pay a lower fraction (45.6%). However, the
differences among the ratios are statistically insignificant.
Another observation that can be learned from Table 6.2 is that contingent pay accounts
for the highest proportion of total compensation and it ranges between 42.9% at the widely-held
firms and 46.8% at the institution-controlled firms. This observation suggests that contingent
compensation have become more and more significant in executive compensation since 1990.
Using TSX data over the period of 1993 to 1995, Zhou (2000) shows that the mean of stock
option related compensation is as high as total cash compensation. Using a Canadian sample
over the period from 1996 to 2002, Chowdhury and Wang (2009) find that the average
percentage of contingent pay to total pay in institution-controlled firms is 50.30%. Therefore, our
results suggest that contingent compensation as a proportion of total compensation have
increased during the 1996-2002 period and then decreased after 2002. It is possible that these
changes are related to the stock market performance during the period of 1996-2000.
We also investigate how the control variables including total assets, ROA, M/B, and D/E
vary among the family-controlled, institution controlled, and widely-held firms. The results are
38
shown in Table 7 from which several observations can be obtained. First, the family-controlled
firms are significantly larger in size and they have significantly higher leverage than the firms in
the institution-controlled or the widely-held firms. However, the sizes of the firms in the last two
groups are not significantly different while widely-held firms are significantly more leveraged
than the institution-controlled firms. In contrast, we find the ROA and M/B of the family-
controlled group are significantly lower than their counterparts of the other two groups. The
institution-controlled firms have similar ROA and M/B as the widely-held firms.
[Insert Table 7 Here]
We analyze the correlations among the different variables in Tables 8.1, 8.2, 8.3, and 8.4.
The correlations among the different variables do not seem to be a problem in our study.
[Insert Table 8.1 Here]
[Insert Table 8.2 Here]
[Insert Table 8.3 Here]
[Insert Table 8.4 Here]
In summary, using unvariate tests we find that both annual bonus and contingent pay are
different among the family-controlled, institution-controlled, and widely-held firms. This means
that the structure of compensation of a firm’s CEO depends on the firm’s ownership structure.
Next, we use multivariate analysis to examine how ownership structure affects CEO
compensation after controlling for other factors that may influence CEO compensation.
39
CHAPTER 6
Multivariate Analysis
Previous empirical studies on CEO compensation mainly concentrate on the relation
between pay and performance. The focus of this thesis is slightly different. We examine the
effect of corporate ownership structure on the levels and structures of the CEO compensation.
Since there is no precedent research on this topic, we borrow and extend the models on pay to
performance relationship to serve our purpose.
Bertrand and Mullaninathan (2001) use the following equation to analyze agency
framework.
𝑦𝑖𝑡 = 𝛽 × 𝑝𝑒𝑟𝑓𝑖𝑡 + 𝛾𝑖 + 𝛿𝑡 + 𝛼𝑥 × 𝑋𝑖𝑡 + εit
Where 𝑦𝑖𝑡 stands for total CEO pay in firm 𝑖 at time t, 𝑝𝑒𝑟𝑓𝑖𝑡 measures firm performance, 𝛾𝑖 are
independent variables that represent the firms fixed variables, 𝛿𝑡 are time fixed variables, and 𝑋𝑖𝑡
are firm characteristics and CEO’s characteristics. The coefficient 𝛽 captures the sensitivity
between performance and CEO pay.
As well, Zhou (2000) estimates the following semi-elasticity specification to examine the
relationship between pay and return and between pay and firm size.
ln (𝐶𝐸𝑂 𝑝𝑎𝑦)𝑡 = 𝑎 + 𝑏 × ln(𝑓𝑖𝑟𝑚 𝑠𝑖𝑧𝑒)𝑡 + 𝑐 × 𝑟𝑒𝑡𝑢𝑟𝑛𝑡
Chowdhury and Wang (2009) used incentive pay parentage and the natural log of
incentive pay as their dependent variables.
We extend the above models by adding a new independent variable, namely ownership
structure, and controlling for the above mentioned pay to performance relationships. Models 1-3
are intended to replicate previous research, particularly Bertrand and Mullaninathan (2001),
while models 4-6 are our main extended models. In model 1, we include the natural logarithm of
total assets (TA), ROA, M/B, D/E, and dummy year variables as our independent variables. Year
2003 serves as the base year and we add the four dummy variables Dyr04, Dyr05, Dyr06, and
Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We use the aggregated sample to
investigate the relationship between the annual bonus and firm characteristics.
𝐿𝑁(𝐶𝑖) = 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀𝐵
+ 𝛽𝐷𝐸
𝐷𝐸
+ ∑𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠 +𝜀𝑖,𝑡 (1)
40
In model 2, we include industry factors by adding three dummy industries to represent
the financial industry (DFin), energy industry (DEgy), and materials industry (DMat). The
resulting equation is:
𝐿𝑁(𝐶𝑖) = 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 +𝛽𝑄𝑀𝐵
+ 𝛽𝐷𝐸
𝐷𝐸
+ �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠
+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 + 𝜀𝑖,𝑡 (2)
In model 3, in order to study the impact of changes in a firm’s CEO, we use a dummy
variable (D1YC) that takes the value of 1 if there is a CEO change in the year (Transient CEO
Group) and 0 otherwise.
𝐿𝑁(𝐶𝑖) = 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀/𝐵 + 𝛽𝐷
𝐸
𝐷𝐸
+ ∑𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠 (3)
+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶 + 𝜀𝑖,𝑡
In model 4, we include variables to control for the concentrated group (firms controlled
either by families or by institutions) and its interaction with each of the independent variables.
The introduction of concentrated ownership dummies change the results reported with model 1.
𝐿𝑁(𝐶𝑖) =
𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀/𝐵 + 𝛽𝐷𝐸
𝐷𝐸
+ �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠
+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 + 𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶 + 𝜔𝐷𝑐𝑜𝑛 𝐷𝑐𝑜𝑛 + 𝜔𝐷𝑐𝑜𝑛𝑅𝑂𝐴𝐷𝑐𝑜𝑛 ∗ 𝑅𝑂𝐴
+𝜔𝐷𝑐𝑜𝑛𝑚𝑏 𝐷𝑐𝑜𝑛 ∗ 𝑀/𝐵 + 𝜔𝐷𝑐𝑜𝑛𝐷/𝐸 𝐷𝑐𝑜𝑛 ∗𝐷𝐸
(4)
In model 5, we separate concentrated-controlled into two groups: family-controlled and
institution-controlled. Including widely-held, we have three variables to control for the
ownership structures. Therefore, we set two dummy variables to compare these three groups. We
compare the annual bonus in the widely-held group to that in both the family-controlled group
and the institution-controlled group, while using widely-held group as the base group.
41
𝐿𝑁(𝐶𝑖) =
𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀/𝐵 + 𝛽𝐷𝐸
𝐷𝐸
+ �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠
(5) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡+𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶 +𝜔𝐷𝐹𝐷𝐹 + 𝜔𝐷𝐹𝑇𝐴 𝐷𝐹 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐹𝑅𝑂𝐴𝐷𝐹 ∗ 𝑅𝑂𝐴 +𝜔𝐷𝐹𝑚𝑏 𝐷𝐹 ∗ 𝑀/𝐵 + 𝜔𝐷𝐹𝐷/𝐸 𝐷𝐹 ∗
𝐷𝐸
+𝜔𝐷𝐼𝐷𝐼 + 𝜔𝐷𝐼𝑇𝐴 𝐷𝐼 ∗ ln(𝑇𝐴)
+𝜔𝐷𝐼𝑅𝑂𝐴𝐷𝐼 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐼𝑚𝑏𝐷𝐼 ∗ 𝑀/𝐵 + 𝜔𝐷𝐼𝐷/𝐸 𝐷𝐼 ∗𝐷𝐸
In order to investigate how annual bonus in family-controlled group is related to annual
bonus in institution-controlled group, we switch our base group, the widely-held group, to the
institution-controlled group, all else equal.
𝐿𝑁(𝐶𝑖) =
𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀/𝐵 + 𝛽𝐷𝐸
𝐷𝐸
+ �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠
(6) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶
+𝜔𝐷𝐹𝑇𝐴 𝐷𝐹 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐹𝑅𝑂𝐴𝐷𝐹 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐹𝑚𝑏𝐷𝐹 ∗ 𝑀/𝐵 + 𝜔𝐷𝐹𝐷/𝐸 𝐷𝐹 ∗𝐷𝐸
+𝜔𝐷𝑊𝑇𝐴 𝐷𝑊 ∗ ln(𝑇𝐴) +𝜔𝐷𝑊𝑅𝑂𝐴𝐷𝑊 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝑊𝑚𝑏𝐷𝑊 ∗𝑀/𝐵 + 𝜔𝐷𝑊𝐷/𝐸 𝐷𝑊 ∗𝐷𝐸
6.1 Natural logarithm of compensation as dependent variable
Table 9.1 reports the results of examining the relation between the annual bonus and
various variables. The table shows that in the absence of control for ownership structure (Models
1, 2, and 3) the annual bonus is positively related to the return on assets and to total assets. When
we divide the sample between concentrated ownership and widely held ownership additional
observations emerge. First, within the widely-held firms the bonus is positively related to total
assets and negatively related to the debt to equity ratio. Also, we find that for the widely held
firms the return on assets is not a significant determinant of the bonus. Second, the concentrated
ownership group seems to pay a higher bonus than the widely held group. Third, within the
concentrated ownership group the bonus is positively and significantly related to the return on
assets, the market to book ratio, and to the debt to equity ratio. This result suggests that the
concentrated ownership firms are linking the higher bonus payments to performance implying
support for our hypothesis that the owners of these firms are monitoring the CEOs more closely.
Fourth, within the concentrated ownership structure the relation between the bonus and asset size
42
is negative and significant which is opposite to the result we found for the widely-held firms.
Again, this result provides support to our hypothesis that the concentrated firms tie their bonuses
to more meaningful measures of performance rather than size.
[Insert Table 9.1 Here]
Table 9.2 reports the results of examining the relation between the annual bonus after
splitting the concentrated ownership group between family-controlled and institution-controlled
firms. The table shows that both family-controlled and institution-controlled firms pay a higher
bonus than the widely held group and that the bonus is positively and significantly related to the
return on assets but negatively related to the size of assets. This result provides support to our
hypothesis that the family-controlled and institution-controlled firms tie their bonuses to more
meaningful measures of performance rather than size. Furthermore, Table 9.2 shows that the
bonus in the family-controlled firms is positively related to the debt to equity ratio but this
relation is not significant for the institution-controlled firms. Finally, Table 9.2 compares the
institution-controlled firms with the family-controlled firms. The results suggest that on average
the two groups pay similar amounts of bonuses but for the family-controlled firms the positive
relation between bonus compensation and the return on assets is stronger. This result suggests
that the level of monitoring by family-controlled firms is stronger which is consistent with our
hypothesis.
[Insert Table 9.2 Here]
Table 10.1 reports the results of examining the relation between the incentive pay and
various variables. The table shows that in the absence of control for ownership structure (Models
1, 2, and 3) the incentive pay is either negatively related to the return on assets or the relationship
is not significant. In contrast, incentive pay seems to be positively and significantly related to
asset size and the market to book ratio. When we divide the sample between concentrated
ownership and widely held ownership (Model 4) additional observations emerge. First, within
the widely-held firms the bonus is positively related to total assets and negatively related to the
return on assets and to the debt to equity ratio. Second, the concentrated ownership group seems
to pay higher incentive compensation but this compensation is negatively related to size while its
relation to return on assets or to the market to book ratio are insignificant.
[Insert Table 10.1 Here]
43
Table 10.2 reports the results of examining the impact of ownership structure on the
incentive pay after splitting the concentrated ownership group between family-controlled and
institution-controlled firms. The table shows that the results of Table 10.1 related to the
concentrated group can be repeated for each of the family-controlled and institution-controlled
firms taken separately. As Model 6 shows, the incentive pay offered by the family-controlled and
the institution-controlled firms does not seem to be related to performance and the levels of
incentive pay provided by the two groups seem to be similar.
[Insert Table 10.2 Here]
Tables 10.1 and 10.2 suggest a significant industry effect in determining incentive
compensation. The energy and materials industries seem to pay significantly higher incentive
compensation than the financial industry firms. This result is consistent regardless of the
ownership structure.
6.2 Bonus and incentive compensation as percentages of total assets
When we analyze the impact of the ownership structure on the annual bonus and the
contingent compensation we find that asset size is a significant determinant of both. This result is
consistent with the findings of previous studies. In this section, we control the impact of asset
size by examining incentive compensation as a ratio of total assets.
In preparation for this analysis, we draw a graph that shows total compensation as a
function of total assets. The graph is shown in Figures 5.1-5.4. Figure 5.1 shows the graph for the
entire dataset while Figures 5.2-5.4 show the relationship for different ranges of assets. The
graphs show that for assets sizes less than $9 billion total compensation is increasing in asset size.
However, at larger asset sizes the graphs do not show a clear pattern of compensation increasing
as a function of total assets.
Figures 6.1 and 6.2 show salary plus bonus compensation as a function of total assets.
Figure 6.1 shows the graph for the entire dataset. This graph shows that bonus plus salary
increases with asset size. Figure 6.2 shows the relation for asset sizes of $20.48 million to $55
billion. For this group of firms, bonus plus salary show an increasing pattern but the rate of
increase will slow down as the size increases.
44
6.2.1. Annual bonus as a percentage of total assets
We use the ratio of annual bonus to total assets (AB-to-TA) as our dependent variable to
test the main hypotheses. Since this dependent variable is restricted at the range of [0, 1], we use
the Tobit model to mitigate the possible problem caused by censored variable. The independent
variables are similar to what included in the previous models. Table 11.1 reports the results.
[Insert Table 11.1 Here]
The table shows that in the absence of control for ownership structure (Models 1, 2, and 3)
the annual bonus as percentage of total assets is positively related to the return on assets and
negatively related to the debt to equity ratio. When we divide the sample between concentrated
ownership and widely held ownership additional observations emerge. First, within the widely-
held firms the bonus continues to be negatively related to the debt to equity ratio but the return
on assets is no longer a significant determinant of the bonus. Second, firms within the
concentrated ownership group seem to pay a lower bonus per dollar of assets than the widely
held group. Third, within the concentrated ownership group the bonus is positively and
significantly related to the debt to equity ratio. Again, this result provides support to our
hypothesis that the concentrated firms tie their bonuses to more meaningful measures of
performance rather than size.
Table 11.2 reports the results of examining the relation between the annual bonus as a
percentage of total assets and the various independent variables after splitting the concentrated
ownership group between family-controlled and institution-controlled firms. The table shows
that both family-controlled and institution-controlled firms pay a lower bonus per dollar of assets
than the widely held group. In addition, within the family-controlled group we find that the
bonus per dollar of assets is positively and significantly related to the return on assets and to the
debt to equity ratio. Furthermore, Model 6 shows that the family-controlled firms pay lower
bonuses per dollar of assets than the institution-controlled firms and that the return on assets is a
positive and significant factor in determining the bonus. The impact of the return on assets is
stronger for the family controlled firms than for the institution controlled firms. This result
provides support to our hypothesis that the family-controlled and institution-controlled firms tie
their bonuses to more meaningful measures of performance rather than size. Furthermore, it
shows that the family-controlled firms provide more monitoring and stronger pay-performance
relation than institution-controlled firms. Finally, Table 11.2 shows that the bonus in the family-
45
controlled firms is positively related to the debt to equity ratio but this relation is negative and
significant for the institution-controlled firms.
[Insert Table 11.2 Here]
Tables 11.1 and 11.2 suggest that new CEOs often receive lower bonus per dollar of
assets than their predecessors. This result is consistent in the various models but it is stronger
when we compare the relations across the various ownership structures. In addition, the tables
show that failing to differentiate between family-controlled and institution-controlled firms may
lead us to conclude that there is a significant industry effect suggesting that financial firms pay
lower bonus per dollar of assets. Controlling for ownership structure shows that this effect is not
significant.
6.2.2. Contingent compensation as a percentage of total assets
We use the ratio of contingent compensation to total assets (CP-to-TA) as our dependent
variable to test the main hypotheses. Since this dependent variable is restricted at the range of [0,
1], we use the Tobit model to mitigate the possible problem caused by censored variable. The
independent variables are similar to what included in the previous models. Table 12.1 reports the
results.
[Insert Table 12.1 Here]
The table shows that in the absence of control for ownership structure (Models 1, 2, and 3)
the contingent compensation as percentage of total assets is negatively related to the return on
assets and to the debt to equity ratio but positively related to the market to book ratio. When we
divide the sample between concentrated ownership and widely held ownership these
observations continue to hold. More important, the table shows that the contingent compensation
as percentage of assets is insignificantly different from the same ratio at the widely-held firms.
Table 12.2 reports the results of examining the relation between the contingent
compensation as percentage of total assets and the various independent variables after splitting
the concentrated ownership group between family-controlled and institution-controlled firms.
The table shows that there is no evidence to suggest that the family-controlled and institution-
controlled firms pay a different percentage of contingent compensation per dollar of assets than
the widely held group.
[Insert Table 12.2 Here]
46
Tables 12.1 and 12.2 show that there is no evidence to suggest that new CEOs receive
different contingent compensation per dollar of assets than their predecessors. This result is
consistent in all the models. In addition, the tables show that there is a strong industry effect.
When we do not control for ownership structure, Table 12.1 shows that contingent compensation
at financial institutions is lower than other industries while the energy and materials sectors pay
greater contingent compensation per dollar of assets than financial institutions or other industries.
Adding control for concentrated ownership does not change the results significantly. However,
when we differentiate between family-controlled and institution-controlled firms, Table 12.2
shows that the results change significantly. Within the family-controlled firms, the debt to equity
ratio seems to have a positive and significant impact on contingent compensation per dollar of
assets. In addition, contingent compensation at financial institutions is no longer significantly
lower than other industries. In contrast, the result related to the energy and materials sectors
continues even after dividing the concentrated ownership firms between family-controlled and
institution-controlled.
6.3 Ownership and the structure of compensation
One objective of this study is to determine whether ownership structure affects the
structure of compensation. In particular, we examine whether bonus payment and contingent
compensation as percentages of total compensation vary across the widely-held, institution-
controlled, and family-controlled firms.
6.3.1 Annual bonus as a percentage of total pay (AB-to-TP)
We use the ratio of bonus compensation to total pay (AB-to-TP) as our dependent
variable to test the main hypotheses. Since this dependent variable is restricted at the range of [0,
1], we use the Tobit model to mitigate the possible problem caused by a censored variable. The
independent variables are similar to those used in the previous models.
Table 13.1 reports the results of models 1-4. It shows that in the absence of controls for
ownership, the annual bonus as proportion of total pay is positively related to the return on assets
and negatively related to the debt to equity ratio. Table 13.2 shows that as we add controls for
ownership, the relation to the return on assets remains unchanged. Furthermore, Model 5 shows
that the ratio of bonus to total compensation in widely-held firms is not significantly different
47
from the same ratio for institution-controlled and family-controlled firms. However, the relation
between the return on assets and the ratio is strongest for the family-controlled firms while there
is no evidence to suggest that the widely-held and the institution-controlled firms display
significantly different relations.
[Insert Table 13.1 Here]
[Insert Table 13.2 Here]
Tables 13.1 and 13.2 also show that new CEOs receive lower proportion of their
compensation in the form of bonus payments. This result is consistent in all the models. In
addition, the tables show that there is a strong industry effect. Table 12.1 shows that when we do
not control for ownership structure, the bonus payments at financial institutions as a percentage
of total pay are higher than other industries while the energy and materials sectors pay less bonus
compensation per dollar of total compensation. As we add controls for ownership structure, the
tables show that there is no evidence to suggest that financial institutions pay significantly
different bonuses than firms in other industries. However, the result related to the energy and
materials sectors continues unchanged after we add control for ownership structure.
6.3.2 Contingent compensation as a percentage of total pay (CP-to-TP)
We use the ratio of contingent compensation to total pay (AB-to-TP) as our dependent
variable to test the main hypotheses. Since this dependent variable is restricted at the range of [0,
1], we use the Tobit model to mitigate the possible problem caused by a censored variable. The
independent variables are similar to those used in the previous models.
Table 14.1 reports the results of models 1-4. It shows that in the absence of controls for
ownership, contingent compensation as proportion of total pay is positively related to the total
assets and the market to book ratio and negatively related to the return on assets. However, the
results of Model 4 show that as we add controls for ownership concentration, the return on assets
becomes insignificant while the debt to equity ratio becomes negatively and significantly related
to the contingent compensation of the widely held firms. Table 14.2 shows that as we
differentiate between family-controlled and institution-controlled firms, we note that both pay
higher proportion of compensation as equity-based. This result is stronger for the institution-
controlled firms. Also, we observe that within the family-controlled firms the ratio of contingent
compensation to total compensation drops with the return on assets.
48
[Insert Table 14.1 Here]
[Insert Table 14.2 Here]
Tables 14.1 and 14.2 also show that new CEOs receive higher proportion of their
compensation in the form of contingent payments. This result is consistent in all the models. In
addition, the tables show that there is a strong industry effect. The contingent payments at
financial institutions as a percentage of total pay are lower than other industries while the energy
and materials sectors pay more contingent compensation per dollar of total compensation. This
result is consistent whether or not we control for ownership structure.
6.4 The relation between compensation and Total Market Return (TMR)
We examine whether the total market return, measured as the capital gains return on the
firm’s common shares plus the dividend yield, affects bonus compensation or stock-based
compensation. We conduct this multivariate analysis without including the market to book ratio.
Table 16 shows that the correlation between M/B and TMR is high.
[Insert Table 16 Here]
The TMR data is obtained from the DataStream database. We start our analysis by using
the basic regression Model 7.11
Model 8 adds industry dummy variables to the basic model.
Model 9 adds a dummy variable to control for CEO departure and replacement, Model 10 adds a
dummy variable to compare the widely-held firms with the concentrated ownership firms. Model
11 adds dummy variables to control for ownership structure and differentiates between widely-
held, family-controlled, and institution-controlled firms. In this model, the widely-held firms
constitute the base group. Model 12 is the same as model 11 except we use the institution-
controlled firms as the base group instead of the widely held.
𝐿𝑁(𝐶𝑖) 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑇𝑀𝑅 + 𝛽𝐷𝐸
𝐷𝐸
+ ∑𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠 +𝜀𝑖,𝑡 (7)
𝐿𝑁(𝐶𝑖) 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑇𝑀𝑅 + 𝛽𝐷
𝐸
𝐷𝐸
+ ∑𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠 (8)
+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 + 𝜀𝑖,𝑡
11 The dependent variable Ci is used to represent compensation and will be used to represent bonus or stock-based
compensation, TMR is total market return, and all the other variables are as defined with Equation 1.
49
𝐿𝑁(𝐶𝑖) 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑇𝑀𝑅 + 𝛽𝐷
𝐸
𝐷𝐸
+ ∑𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠 (9)
+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶 + 𝜀𝑖,𝑡
𝐿𝑁(𝐶𝑖)
𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑇𝑀𝑅 + 𝛽𝐷𝐸
𝐷𝐸 + �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠
(10) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶
+𝜔𝐷𝑐𝑜𝑛 𝐷𝑐𝑜𝑛 ∗ ln(𝑇𝐴) + 𝜔𝐷𝑐𝑜𝑛𝑅𝑂𝐴𝐷𝑐𝑜𝑛 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝑐𝑜𝑛𝑚𝑏 𝐷𝑐𝑜𝑛 ∗ 𝑇𝑀𝑅 + 𝜔𝐷𝑐𝑜𝑛𝐷/𝐸 𝐷𝑐𝑜𝑛 ∗𝐷𝐸 + 𝜀𝑖,𝑡
𝐿𝑁(𝐶𝑖)
𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀/𝐵 + 𝛽𝐷𝐸
𝐷𝐸 + �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠
(11) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝑇𝑀𝑅𝑇𝑀𝑅 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶
+𝜔𝐷𝐹𝐷𝐹 + 𝜔𝐷𝐹𝑇𝐴 𝐷𝐹 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐹𝑅𝑂𝐴𝐷𝐹 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐹𝑚𝑏 𝐷𝐹 ∗ 𝑀/𝐵 + 𝜔𝐷𝐹𝐷/𝐸 𝐷𝐹 ∗𝐷𝐸
+𝜔𝐷𝐼𝐷𝐼 + 𝜔𝐷𝐼𝑇𝐴 𝐷𝐼 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐼𝑅𝑂𝐴𝐷𝐼 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐼𝑚𝑏𝐷𝐼 ∗ 𝑀/𝐵 + 𝜔𝐷𝐼𝐷/𝐸 𝐷𝐼 ∗𝐷𝐸
+𝜀𝑖,𝑡
𝐿𝑁(𝐶𝑖)
𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀/𝐵 + 𝛽𝐷𝐸
𝐷𝐸 + �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠
(12) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝑇𝑀𝑅𝑇𝑀𝑅 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶
+𝜔𝐷𝐹𝑇𝐴 𝐷𝐹 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐹𝑅𝑂𝐴𝐷𝐹 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐹𝑚𝑏𝐷𝐹 ∗ 𝑀/𝐵 + 𝜔𝐷𝐹𝐷/𝐸 𝐷𝐹 ∗𝐷𝐸
+𝜔𝐷𝑊𝑇𝐴 𝐷𝑊 ∗ ln(𝑇𝐴) +𝜔𝐷𝑊𝑅𝑂𝐴𝐷𝑊 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝑊𝑚𝑏𝐷𝑊 ∗𝑀/𝐵 + 𝜔𝐷𝑊𝐷/𝐸 𝐷𝑊 ∗𝐷𝐸 + 𝜀𝑖,𝑡
The results of analysing the relation between bonus compensation and TMR are reported
in Tables 17.1 and 17.2. The table shows that TMR is not a significant factor in determining
bonus compensation. At the same time, the relation between bonus compensation and the other
decision and control variables remains unchanged in terms of the direction and significance. In
addition, this observation can be repeated after controlling for industry effects, CEO turnover,
and ownership structure.
[Insert Table 17.1 Here]
[Insert Table 17.2 Here]
50
Similarly, we examine the relation between contingent pay and TMR. The results are
reported in Tables 18.1 and 18.2. TMR does not have a significant impact on contingent pay. At
the same time, the relation between stock-based compensation and the other decision and control
variables remains unchanged in terms of the direction and significance. Adding controls for
industry effects, CEO turnover, and ownership structure does not change the insignificant
relation between stock-based compensation and total market return.
[Insert Table 18.1 Here]
[Insert Table 18.2 Here]
51
CHAPTER 7
Conclusions and Recommendations for Further Research
This study examines the role that ownership structure plays in the governance of large,
publicly traded firms in Canada. This subject is important given the steady rise in the levels of
CEO compensation, particularly contingent compensation, over the past two decades. We argue
that the split of CEO compensation between salary, bonus, stock-based, and other compensation
would be different across different ownership structures.
Overall, our analysis leads to many observations related to the relation between
ownership structure and compensation. First, both family-controlled and institution-controlled
firms pay higher bonus than widely-held firms and that the bonus in concentrated ownership
firms is positively and significantly related to the return on assets. This result provides support to
our hypothesis that the family-controlled and institution-controlled firms tie their bonuses to
meaningful measures of performance. Second, the analysis suggests that on average the family-
controlled and institution-controlled firms pay similar amounts of bonuses but for the family-
controlled firms the positive relation between bonus compensation and the return on assets is
stronger. This result suggests that the level of monitoring by family-controlled firms is stronger
which is consistent with our hypothesis.
We examine the relation between ownership and contingent compensation. We observe
that within the widely-held firms contingent compensation is positively related to total assets and
negatively related to the return on assets and to the debt to equity ratio. Second, the concentrated
ownership group seems to pay higher incentive compensation but this compensation is
negatively related to size while its relation to return on assets or to the market to book ratio are
insignificant. Splitting the concentrated ownership group between family-controlled and
institution-controlled firms reinforces the results reported for the concentrated ownership group.
In conclusion, the incentive pay offered by the family-controlled and the institution-controlled
firms does not seem to be related to performance and the levels of incentive pay provided by the
two groups seem to be similar.
We examine the impact of ownership structure on the compensation per dollar of assets.
Several observations emerge. First, both family-controlled and institution-controlled firms pay a
lower bonus per dollar of assets than the widely held group. Second, within the family-controlled
group we find that the bonus per dollar of assets is positively and significantly related to the
52
return on assets. Third, the family-controlled firms pay lower bonuses per dollar of assets than
the institution-controlled firms and that the return on assets is a positive and significant factor in
determining the difference in bonus. Fourth, the impact of the return on assets is stronger for the
family controlled firms than for the institution controlled firms. This result provides support to
our hypothesis that the family-controlled and institution-controlled firms tie their bonuses to
measurable metrics of performance such as return on assets. In addition, the results show that the
family-controlled firms provide more monitoring and exhibit stronger pay-performance relation
than institution-controlled firms.
Similarly, we examine the impact of ownership structure on the ratio of contingent
compensation per dollar of assets. We find that the ratio is negatively related to the return on
assets but positively related to the market to book ratio. At concentrated ownership firms,
contingent compensation as percentage of assets is insignificantly different from the same ratio
at the widely-held firms. In addition, we find no evidence to suggest that the family-controlled
and institution-controlled firms pay a different percentage of contingent compensation per dollar
of assets than the widely held group. These results suggest that stock-based compensation is not
a major tool to control the behaviour of managers but it is offered as part of a competitive
compensation package consistent with market practices.
Additional support to this conclusion is obtained when we examine the relation between
ownership structure and the ratio of bonus compensation to total compensation. The results show
that the ratio in widely-held firms is not significantly different from the same ratio at institution-
controlled and family-controlled firms. However, the relation between the return on assets and
the ratio is strongest for the family-controlled firms. In contrast, there is no evidence to suggest
that the widely-held and the institution-controlled firms display significantly different relations.
Similarly, we examine the impact of the ownership structure on the ratio of contingent
compensation to total compensation. The results show that in comparison to widely-held firms
both family-controlled and institution-controlled firms pay higher proportion of compensation as
equity-based. This result is stronger for the institution-controlled firms. Also, we observe that
within the family-controlled firms the ratio of contingent compensation to total compensation
drops with the return on assets.
53
Our results suggest a significant industry effect in determining bonus and stock-based
compensation. We find that the financial industry pays less bonus or contingent compensation
than the other industries. At the same time, the energy and materials industries seem to pay
significantly higher bonus and contingent compensation than the financial industry. When we
scale compensation by total assets, we find no differences in the bonus per dollar of assets
among the industry groups. In contrast, we find that contingent compensation per dollar of assets
at financial institutions is insignificantly lower than other industries while the energy and
materials sectors pay significantly higher contingent compensation per dollar of assets than
financial institutions or other industries.
Our analysis also compares the compensation of incoming CEOs with the compensation
of their predecessors. The results suggest that incoming CEOs often receive lower bonus per
dollar of assets than their predecessors. Consistent with these results, we also find that incoming
CEOs also receive lower proportion of their compensation in the form of bonus payments. In
contrast, the analysis shows that incoming CEOs receive higher proportion of their compensation
in the form of contingent payments.
In summary, we make significant contributions to the existing literature by comparing
bonus and stock-based compensation across widely-held, institution-controlled, and family-
controlled firms. Moreover, we compare and analyze the difference between incoming CEO
compensation and outgoing CEO compensation across the three ownership structures. Overall,
we obtain several important conclusions. First, we find that the ownership structure is very
important in determining the level and structure of CEO compensation. In particular, family-
controlled and institution-controlled firms have significantly different compensation structures.
Second, we find that bonus compensation is more associated with performance than stock-based
compensation. Third, control by a family seems to provide more effective and meaningful
monitoring of managers than control by institutions. Fourth, the structure of compensation
provided to incoming CEOs is significantly different from the structure of compensation
provided to outgoing CEOs. In particular, incoming CEOs seem to obtain less bonus
compensation and more stock-based compensation than their predecessors.
Our findings are useful for investors, academics, policy makers, and regulators. All these
stakeholders should be aware that there is room for improvement in the relationship between
managers and shareholders. Institutions do not seem to be providing active monitoring of
54
managers. Therefore, the existence of an institutional block holder does not necessarily reduce
agency costs. Control by a family seems to provide better monitoring but as documented by the
literature it has agency costs of its own. Therefore, there is need for improvement in corporate
governance practices. In addition, our results show that the compensation package is a significant
tool for influencing the CEO. However, the bonus component seems to be the only form of
compensation that is associated with performance which suggests that stock-based compensation
is not accomplishing its objectives of aligning the interests of managers and shareholders. We
speculate that two factors may be magnifying the problem. First, there is no accurate system that
sets the optimal stock-based compensation. Academics and practitioners should focus their
attention to develop such a system. Second, there is lack of information regarding the ultimate
value of stock-based compensation. The recent regulations which require the disclosure of the
value attached to stock-based compensation is a good step in this direction. We suggest that
policy makers and regulators should encourage better accounting for stock-based compensation.
This study has limitations. First, prior to 2005 Canadian firms were not required to
provide an estimate of the full monetary value of CEO stock-based compensation in their
financial statements. Thus, for some observations we use the approximate value estimated using
the Black and Scholes (1972) option pricing model. The approximations may introduce some
errors. Second, our independent variables do not include CEO characteristics, which, according
to previous studies, might have an impact on CEO compensation.
55
References
Amoako-Adu, B., Smith, B. F., 2001, Dual Class Firms: Capitalization, Ownership Structure,
and Recapitalization Back into Single Class, Journal of Banking and Finance, Vol.25, pp.
1083–1111.
Anderson, R., Reeb, D., 2003, Founding-Family Ownership and Firm Performance: Evidence
from the S&P 500, Journal of Finance, Vol.58, pp.1301–1327.
Ang, Janmes S., Cole, Rebel A., Lin, Janmes Wuh., 2002, Agency Costs and Ownership
Structure, Journal of Finance, Vol. 55, Issue 1 (Dec), pp. 81-106.
Arya, Avinash, Huey-Lian, Sun, 2004, Stock Option Repricing: Heads I Win, Tails You
Lose, Journal of Business Ethics, Vol. 50, Issue 4 (Apr), pp. 297-312.
Barenbaum, Les., and Schubert, Walt., 1993, Measuring the Value of Executive Stock Options,
Compensation & Benefits, Vol. 25 (Oct), pp. 19-24.
Baumol, W., 1959, Business Behavior, Value and Growth. New York.
Bebchuk, L. and Stole L., 1993, Do short-term objectives lead to under-or overinvestment in
contingent projects? Journal of finance, Vol. 48, pp. 719-730.
Ben-Amar, Walid. and Andre, Paul., 2006, Separation of Ownership from Control and Acquiring
Firm Performance: The Case of Family Ownership in Canada, Journal of Business
Finance & Accounting, Vol.33, No.3&4, pp.517–543.
Bennedsen, Morten et al., 2007, Inside the Family Firm: the Role of Families in Succession
Decisions and Performance, The Quarterly Journal of Economics, Vol. 122, No. 2, pp.
647-691.
Berle, Adolf A., Jr. and Gardner C, Means., 1932, The Modern Corporation and Private Property
(Macmillan, New York, NY).
Chourou, Lamia, 2010, Compensation of Owner Managers in Canadian Family-Owned
Businesses: Expropriation of Minority Shareholders, Canadian Journal of Administrative
Sciences, Vol. 27, pp. 95-106.
Chowdhury, S.D., Wang, E.Z., 2009, Institutional Activism Types and CEO Compensation: A
Time-Series Analysis of Large Canadian Corporation, Journal of Management, Vol. 35,
pp. 5-36.
56
Chrisman, James J., Chua, Jess H., and Litz, Reginald, A., 2004, Comparing the Agency Costs of
Family and Non-Family Firms: Conceptual Issues and Exploratory Evidence,
Entrepreneurship Theory and Practice, Vol. 28, Issue. 4 (May), pp. 335-354.
Ciscel, D. and Carroll T., 1980, The Determinants of Executive Salaries: An Econometric Survey,
Review of Economics and Statistics, Vol.62, No.1, pp. 7-13.
Claessens, S., Djankov, S. and Lang, L., 2000. The Separation of Ownership and Control in East
Asian Corporations, Journal of Financial Economics, Vol.58, pp. 81–112.
Daily, C. M., Johnson, J. L., Ellstrand, A. E., and Dalton, D. R., 1998, Compensation Committee
Composition as a Determinant of CEO Compensation, Academy of Management Journal,
Vol.41, pp. 209-220.
David, P., Kochhar, R., and Levitas, E., 1998, The Effect of Institutional Investors on the Level
and Mix of CEO Compensation, Academy of Management Journal, Vol.41, No.2, pp. 200.
Demsetz, H., 1983, The Structure of Ownership and the Theory of the Firm, Journal of Law and
Economic, Vol.26, pp. 375-390.
Demsetz, H., Lehn, K., 1985, The Structure of Corporate Ownership: Causes and Consequences.
Journal of Political Economy, Vol.93, pp. 1155–1177.
Faccio, Maria, and Lang, Larry P.H., 2002, The Ultimate Ownership of Western European
Corporations, Journal of Financial Economics, Vol.65, pp. 365-395
Fahlenbrach, Budiger, Stulz, Rene M., 2011, Bank CEO Incentives and the Credit Crisis, Journal
of Financial Economics, Vol.99, pp.11-26.
Fama, Eugene F., and Jensen, Michael C., 1983, Separation of Ownership and Control. Journal
of Law & Economics, Vol.26, pp. 301-325.
Fifth, Michael., Fung, Peter M.Y., Rui, Oliver M., 2006, Firm Performance, Governance
Structure, and Top Management Turnover in a Transitional Economy, Journal of
Management Studies, Vol. 43, Issue. 6 (May), pp. 1289-1330.
Florackie, C. and Ozkan, A., 2004, Agency Cost and Corporate Governance Mechanisms:
Evidence for UK firms, paper presented at European Finance Association Meeting
Gibbons, R. and Murphy, K. J., 1990, Relative Performance Evaluation for Chief Executive
Officers, Industrial and Labor Relations Review, Vol.43, No.3 (Feb), pp. 30-51.
Gray, S. R., and Cannella, A. A., 1997, The Role of Risk in Executive Compensation, Journal of
Management, Vol.23, pp. 57-540.
57
Grossman, Sanford, and Hart, Oliver., 1988, One Share-one Vote and the Market for Corporate
Control, Journal of Financial Economics, Vol.20, pp.175-202
Guidry, Flora, Leone, Andrew J., Rock, Steve, 1999, Earnings-based Bonus Plans and Earnings
Management by Business-unit Managers, Journal of Accounting and Ecomomics, Vol. 26,
pp.113-142.
Hall, Brian J., Liebman, Jeffrey B.,1998, Are CEOS Really Paid Like Bureaucrats? The
Quarterly Journal of Economics, Vol.113, Issue. 3 (Aug), pp. 653-691.
Hart, Oliver D., 1983, The Market Mechanism as an Incentive Scheme, Bell Journal of
Economics, Vol.14, pp.366-382
Hartzell, Jay C., and Starks, Laura T., 2003, Institutional Investors and Executive Compensation,
The Journal of Finance, Vol.9, No.6, pp.2351-2374.
Healy, Joanne P., and Cole, Elizabeth T., 2000, The Effect of Corporate Owneship Structure on
CEO Compensation, Journal of Deferred Compensation, Vol.3, No.2, pp. 61-72
Huson, Mark R., Parrino, Robert., and Starks, Laura T., 2002, Internal Monitoring Mechanisms
and CEO Turnover: A Long-Term Perspective, Journal of Finance, Vol. 56, Issue.6 (Dec),
pp. 2265-2297.
Jacobs, M., 1991, Short-term America: the Causes and Cures of Our Business Myopia, Harvard
Business School Press.
Jensen, Michael C., and Richard S. Ruback, 1983, The Market for Corporate Control: The
Scientific Evidence, Journal of Financial Economics, Vol.11, pp.5-50.
Jensen, Michael C., and William H. Meckling, 1976, Theory of the Firm: Managerial Behavior,
Agency Costs and Ownership Structure, Journal of Financial Economics, Vol.3, No. 4
pp.305-360.
Jensen, Michael C., and K. Murphy, 1990, Performance Pay and Top Management Incentives.
Journal of Political Economy, Vol. 98, pp. 225-264
Jeppson, Catherine T., Smith, Wayne W., Stone, Ronald S., 2009, CEO Compensation and Firm
Performance: Is There Any Relationship, Journal of Business & Economics Research, Vol.
7, No.11(Nov), pp. 81-93.
Kaplan, Steven N., 1994, Top Executive Rewards and Firm Performance: A Comparison of
Japan and the United States, The Journal of Political Economy, Vol. 102, Issue. 3 (Jun),
pp. 510-546
58
Kaplan, Steven N., and Minton, Bernadette., 2006, How Has CEO Turnover Changed?
Increasingly Performance Sensitive Boards and Increasingly Unesay CEOs, Working
paper, University of Chicago
Kato, Takao and Long, Cheryl., 2006, Executive Compensation, Firm Performance, and
Corporate Governance in China: Evidence from Firms Listed in the Shanghai and
Shenzhen Stock Exchanges, Economic Development and Cultural Change, Vol.54, No.4,
pp. 945-83.
Klein, Peter C., Daniel M. Shapiro, and Jeffrey Young, 2005, Corporate Governance, Family
Ownership and Firm Value: The Canadian Evidence, Corporate Governance: An
International Review, Vol. 13, No. 6, pp. 769-784.
Kole, Stacey R., 1997, The Complexity of Compensation Contracts, Journal of Financial
Economicas, Vol. 43, pp. 79-104.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R., 1998, Law and Finance, Journal
of Political Economy, Vol.106, pp. 1113–1155.
La Porta, R., Lopez-de-Silanes, F., Shleifer, F. and Vishny, R., 1999, Corporate Ownership
around the World, Journal of Finance, Vol.54, pp. 471–517.
Lambert, Richard A., Larcker, David F., and Verrecchia, Robert E., 1991, Portfolio
Cosiderations in Valuing Executive Compensation, Journal of Accounting Research, Vol.
29, No. 1 (Spring), pp. 129-149.
Leonard, J., 1990, Executive Pay and Firm Performance, Industrial and Labor Relations Review,
Vol.43, No.3, pp. 13-29.
Lewellen, W. and Huntsman, B. 1970, Managerial Pay and Corporate Performance, American
Economic Review, Vol.60, No.4, pp. 710-20.
Mehran, Hamid., 1995, Executive Compensation Structure, Ownership, and Firm Performance,
Journal of Financial Economics, Vol. 38, pp. 163-184.
Morck, Randall., Shleifer, Andrei., and Vishny, Robert W., 1988, Management Ownership and
Market Valuation: An Empirical Analysis, Journal of Financial Economics, Vol. 20 (Jan-
Mar), pp. 293-315.
Murphy, K. J., 1985, Corporate Performance and Managerial Remuneration: An Empirical
Analysis, Journal of Accounting and Economics, Vol.7, No.1-3, pp. 11-42.
59
Murphy, K. J., 1986, Incentives, Learning, and Compensation: a Theoretical and Empirical
Investigation of Managerial Labor Contracts, Rand Journal of Economics, Vol.17, No.1,
pp. 59-76.
Murphy, Kevin J.,1999, Executive Compensation, Hand book of Labor Economics, Vol.3, No. 2,
pp. 2485-2563.
Ozkan, Neslihan., 2007, Do Corporate Governance Mechanisms Influence CEO Compensation?
An empirical investigation of UK companies, Journal of Multinational Financial
Managerment, Vol. 17, Issue 5 (Dec.), pp. 349-364.
Roe, S. and Lee-Sing, C. R., 1996, Governance Structure, Corporate Decision-Making and Firm
Performance in North America. Working Paper, No.7, Industry Canada.
Schulze, W. S., Lubatkin, M. H., and Dino, R. N., 2003, Exploring the Agency Consequences of
Ownership Dispersion among the Directors of Private Family Firms. Academy of
Management Journal, Vol.46, No.2, pp. 179–194.
Shavell, Steven, 1979, Risk Sharing and Incentives in the Principal and Agent Relationship, The
Bell Journal of Economics, Vol.10, No.1, pp.55-73.
Shaw, Kenneth W., Zhang, May H., 2010, Is CEO Cash Compensation Punished for Poor Firm
Performance? The Accounting Review, Vol.85, Issue.3, pp. 1065-1093.
Shleifer, A and Vishny, R.W., 1986, Large Shareholders and Corporate Control. The Journal of
Political Economy, Vol.94, No.3, pp. 461-488.
Shleifer, A. and Vishny, R., 1997, A Survey of Corporate Governance, Journal of Finance, Vol
52, pp. 737-783
Singh, M. and Davidson, W. N., 2003, Agency Costs, Ownership Structure and Corporate
Governance Mechanism. Journal of Banking and Finance, Vol.27, pp. 793–816.
Smith, Michael., 1996, Shareholder Activism by Institutional Investors: Evidence from CalPERS,
Journal of Finance, Vol.51, pp. 227-252.
Tannous, George F., and Bin Cheng, (2007), Canadian Takeover Announcements and the Job
Security of Top Managers, Canadian Journal of Administrative Sciences, Volume 24,
Issue 4 (December), Pages 250-267.
Villalonga, Amit, 2006, How do Family Ownership, Control and Management Affect Firm
Value? Journal of Financial Economics, Vol.80, pp. 385-517.
60
Warner, J., R. Watts and Wruck, K., 1988, Stock Prices and Top Management Changes, Journal
of Financial Economics, Vol.20, pp. 461-92.
Zhou, Xianming., 2000, CEO Pay, Firm Size, and Corporate Performance: Evidence from
Canada, Canadian Journal of Economics, Vol.33, pp. 213–51.
61
Appendix I: Tables Table 1.1: Mean value of CEO monetary compensation in permanent CEOs firms: comparison between family controlled, institution-controlled, and widely-held firms 2003 2004 2005 2006 2007 Panel A: Family-Controlled Salary 779,140 812,010 854,452 892,431 962,269 Annual Bonus 912,965 1,700,900 1,769,673 2,330,298 2,643,697 Contingent Pay 2,758,991 3,767,955 4,406,136 4,024,199 4,477,935 Total Compensation 4,712,431 6,391,648 7,319,985 7,499,857 8,365,905 2003 2004 2005 2006 2007 Panel B: Institution-Controlled Salary 479,306 515,045 577,061 666,922 775,601 Annual Bonus 435,972 574,395 620,760 801,291 1,033,146 Contingent Pay 1,471,428 1,599,960 2,038,916 3,428,518 3,137,185 Total Compensation 2,545,116 2,888,358 3,646,281 5,031,109 5,148,305 2003 2004 2005 2006 2007 Panel C: Widely-Held Salary 487,555 530,926 552,993 580,543 607,251 Annual Bonus 599,530 694,278 869,917 1,030,685 946,963 Contingent Pay 2,128,817 2,061,454 2,242,162 2,116,396 2,133,368 Total Compensation 3,328,682 3,494,418 3,895,404 4,040,169 3,871,169 Table 1.2: Mean value of salary, annual bonus, and contingent pay as a percentage of total pay in permanent CEOs firms: comparison between family controlled, institution-controlled, and widely-held firms 2003 2004 2005 2006 2007 Panel A: Family-Controlled S-to-TP 0.1653 0.1270 0.1167 0.1190 0.1150 AB-to-TP 0.1937 0.2661 0.2418 0.3107 0.3160 CP-to-TP 0.5855 0.5895 0.6019 0.5366 0.5353 2003 2004 2005 2006 2007 Panel B: Institution-Controlled S-to-TP 0.1883 0.1783 0.1583 0.1326 0.1507 AB-to-TP 0.1713 0.1989 0.1702 0.1593 0.2007 CP-to-TP 0.5781 0.5539 0.5592 0.6815 0.6094 2003 2004 2005 2006 2007 Panel C: Widely-Held S-to-TP 0.1465 0.1519 0.1420 0.1437 0.1569 AB-to-TP 0.1801 0.1987 0.2233 0.2551 0.2446 CP-to-TP 0.6395 0.5899 0.5756 0.5238 0.5511 Notes: S-to-TP denotes salary as a percentage of total pay, AB-to-TP denotes annual bonus as a percentage of total pay, and CP-to-TP denotes contingent compensation as a percentage of total pay.
62
Table 2.1: Descriptive statistics in the transient CEO group Mean s.d. Min Med Max Mode New CEOs Salary 678,705 388,480 0 586,446 62,288,180 933,333
Annual Bonus 868,823 1,626,077 0 459,776 9,093,013 831,600
Contingent Pay 3,905,164 6,574,777 0 1,665,315 31,907,517 0
Total Pay 5,785,076 7,464,049 873,104 2,937,326 35,165,720 N/A
Old CEOs
Salary 784,580 629,311 110,500 575,000 3,450,000 600,000
Annual Bonus 652,871 993,679 0 275,252 4,816,720 0
Contingent Pay 2,459,336 5,941,128 0 637,712 31,334,307 0
Total Pay 4,194,576 6,953,677 311,886 2,329,344 37,076,147 N/A
Observations 31
Note: 1. ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. 2. Data sample include transient CEO firms without changing ownership structures. 3. AB-to-TP denotes annual bonus as a percentage of total pay and CP-to-TP denotes
contingent compensation as a percentage of total pay.
63
Table 2.2: Comparison of the compensation of incoming CEOs and that of their predecessors Panel A: total sample T-test: two sample assuming unequal variances DIM t-stats p-value Salary -105875 -0.797 0.215 Annual Bonus 260856 0.762 0.225 Contingent Compensation 1445827 0.908 0.184 Total 1590500 0.868 0.194 Z-test: two sample for means DIM z-stats p-value S-to-TP -0.073 -1.692** 0.045 AB-to-TP -0.011 -0.212 0.416 CP-to-TP 0.109 1.532* 0.063 Panel B: family group T-test: two sample assuming unequal variances DIM t-stats p-value Salary -388529 -0.824 0.215 Annual Bonus 254098 0.776 0.228 Contingent Compensation 4746786 0.760 0.231 Total 4481346 0.653 0.263 Z-test: two sample for means DIM z-stats p-value S-to-TP -0.142 -1.310 0.095 AB-to-TP 0.033 0.350 0.364 CP-to-TP 0.210 1.074 0.141 Panel C: Institution group T-test: two sample assuming unequal variances DIM t-stats p-value Salary -17245 0.151 0.441 Annual Bonus 153785 0.425 0.338 Contingent Compensation 166215 0.738 0.235 Total 1615033 1.248 0.116 Z-test: two sample for means DIM z-stats p-value S-to-TP -0.048 -0.766 0.222 AB-to-TP -0.037 -0.402 0.344 CP-to-TP 0.131 1.524* 0.064 Panel D: Widely-held group T-test: two sample assuming unequal variances DIM t-stats p-value Salary -30748 -0.248 0.404 Annual Bonus -49925 -0.181 0.30 Contingent Compensation -456202 -0.323 0.375 Total -278123 -0.145 0.443 Z-test: two sample for means DIM z-stats p-value S-to-TP -0.058 -0.792 0.214 AB-to-TP -0.008 -0.108 0.457 CP-to-TP 0.002 0.145 0.442 Note: 1. ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. 2. Data sample include transient CEO firms without changing ownership structures. 3. DIM denotes differences in mean.
64
Table 3: Descriptive statistic related to the dependent and control variables Mean s.d. Min Med Max Mode Dependent Variables Annual Bonus 913,019 1,465,376 0 500,000 12,929,728 0
Contingent Pay 2,885,466 5,149,953 0 1,206,930 63,815,034 0
Control Variables
ln(TA) 21.77 1.74 16.83 21.56 27.12 21.61
ROA 4.17 7.13 -44.97 3.70 32.02 1.63
M/B 2.73 1.71 0.55 2.35 16.49 1.89
D/E (%) 85.77 135.43 0 44.93 1383.31 0
Notes: 1. Ln (TA) denotes capital expenditure, ROA denotes return on assets, M/B denotes
market-to-book value, and D/E denotes debt to equity ratio. 2. Data include both permanent CEO firms and transient CEO firms
65
Table 4: Descriptive statistic related to the dependent and control Variables in three different ownership structures
Panel A: Family-controlled group Dependent Variables Mean s.d. Min Med Max Mode Annual Bonus 1,460,940 2,431,252 0 748,750 12,929,728 0 Contingent pay 5,489,968 9,914,024 0 1,326,253 63,815,034 0 Control Variables Ln(TA) 22.76 1.41 20.47 22.59 25.61 N/A ROA 2.54 4.61 -17.29 1.90 12.27 N/A M/B 2.29 1.21 0.55 1.89 7.68 1.89 D/E (%) 139.76 241.91 0 53.45 1383.31 0 Panel B: Institution-controlled group Dependent Variables Mean s.d. Min Med Max Mode Annual Bonus 824,449 1,317,417 0 475,000 12,807,300 0 Contingent pay 2,423,975 3,312,135 0 1,287,596 27,404,855 0 Control Variables ln(TA) 21.41 1.37 16.83 21.41 24.73 N/A ROA 4.93 7.91 -41.91 4.18 32.02 N/A M/B 2.89 1.84 0.61 2.36 12.43 1.51 D/E (%) 65.66 72.80 0 47.17 437.66 0 Panel C: Widely-Held Group Dependent Variables Mean s.d. Min Med Max Mode Annual Bonus 775,703 944,806 0 476,937 5,000,000 0 Contingent pay 2,271,854 3,066,349 0 1,087,309 17,832,455 0 Control Variables ln(TA) 21.65 2.01 17.65 21.24 27.12 21.61 ROA 4.10 7.07 -44.97 3.95 23.48 -0.31 M/B 2.76 1.73 0.60 2.40 16.49 1.60 D/E (%) 84.21 115.01 0 40.99 991.55 0 Notes: 1. Ln (TA) denotes capital expenditure, ROA denotes return on assets, M/B denotes
market-to-book ratio, and D/E denotes debt to equity ratio. 2. Data include both permanent CEO firms and transient CEO firms. Observations are 96
in family-controlled group, 228 in institution-controlled group, and 236 in widely held group.
66
Table 5: Descriptive statistic related to the dependent and control variables in different industries Panel A: Financial Industry Dependent Variables Mean s.d. Min Med Max Mode Bonus pay 1,554,841 1,484,468 0 1,400,000 5,000,000 0 Contingent pay 4,402,151 5,583,377 0 1,935,630 20,479,090 0 Control Variables Ln(TA) 23.89 2.16 18.87 23.69 27.12 N/A ROA 2.34 2.73 -1.98 1.41 12.08 1.02 M/B 2.50 1.13 0.76 2.42 6.98 0.76 D/E (%) 114.55 134.36 0 47.68 480.70 0 Panel B: Energy Industry Dependent Variables Mean s.d. Min Med Max Mode Bonus pay 371,484 488,355 0 287,500 3,874,900 0 Contingent pay 1,966,984 2,468,270 0 883,167 11,711,567 0 Control Variables ln(TA) 20.91 1.33 17.65 20.85 23.62 N/A ROA 6.31 7.26 -11.17 5.39 32.02 N/A M/B 2.66 1.25 0.64 2.36 6.39 N/A D/E (%) 89.81 92.15 0 59.33 437.66 0 Panel C: Material Industry Dependent Variables Mean s.d. Min Med Max Mode Bonus pay 690,189 770,581 0 513,500 4,822,429 0 Contingent pay 2,446,107 2,653,612 0 1,556,225 10,382,308 0 Control Variables ln(TA) 21.21 1.17 18.42 21.23 23.94 N/A ROA 5.97 7.11 -12.72 5.04 30.74 N/A M/B 3.06 1.90 0.60 2.49 9.34 3.19 D/E (%) 40.40 50.15 0 24.80 300.16 0 Panel D: Other Industries Dependent Variables Mean s.d. Min Med Max Mode Bonus pay 970,372 1,775,991 0 548,000 12,929,728 0 Contingent pay 2,865,062 6,228,979 0 1,168,075 63,815,034 0 Control Variables ln(TA) 21.57 1.29 16.83 21.53 25.51 N/A ROA 3.34 7.77 -44.97 3.98 24.10 1.63 M/B 2.70 1.90 0.55 2.24 16.49 2.22 D/E (%) 93.32 164.71 0 47.04 1383.31 0 Notes: 1. Ln (TA) denotes capital expenditure, ROA denotes return on assets, M/B denotes
market-to-book value, and D/E denotes debt to equity ratio. 2. Data include both permanent CEO firms and transient CEO firms. Observations are 90
in the financial industry, 90 in the energy industry, 110 in the material industry, and 270 in the other industries.
67
Table 6.1: T-test: two sample assuming unequal variances (monetary terms across three ownerships) Family/Institution Family/Widely Held Institution/Widely Held Annual Bonus 2.420** 2.680*** 0.457 Contingent Pay 2.961*** 3.120*** 0.513 Table 6.2: Descriptive statistics mean value annual bonus and contingent pay Panel A: Mean Family Institution Widely Held AB-to-TP 0.236 0.216 0.219 CP-to-TP 0.456 0.468 0.429 Panel B: z-test: Family/Institution Family/Widely Held Institution/Widely Held AB-to-TP 0.789 0.680 -0.158 CP-to-TP -0.334 0.712 1.442 Note: 1. ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. 2. Z-test: two Sample for Means (different ownership structures) 3. Data include both permanent CEO firms and transient CEO firms 4. AB-to-TP denotes annual bonus as a percentage of total pay and CP-to-TP denotes
contingent compensation as a percentage of total pay.
68
Table 7: Descriptive statistics mean value for the control variables Family/Institution Family/Widely Held Institution/Widely Held Ln(TA) 7.595*** 5.679*** -1.172 ROA -3.406*** -2.38*** 1.191 M/B -3.449*** -2.819*** 0.770 D/E 2.907*** 2.153** 2.200** Note: 1. ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. 2. Z-test: two Sample for Means (different ownership structures) 3. Data include both permanent CEO firms and transient CEO firms
69
Table 8.1: Correlation of variables in the aggressive sample Panel A ln (AB) ln (CP) ln (TA) ROA M/B D/E ln(AB) 1 ln(CP) 0.2139 1 ln(TA) 0.1428 0.2056 1 ROA 0.2371 -0.0419 -0.0642 1 M/B 0.0547 0.0629 -0.0919 0.0944 1 D/E -0.0687 0.0563 0.3141 -0.1600 0.0120 1 Panel B AB-to-TP CP-to-TP ln(TA) ROA M/B D/E AB-to-TP 1 CP-to-TP -0.5885 1 ln(TA) -0.0167 0.1723 1 ROA 0.1830 -0.0651 -0.0642 1 M/B -0.0312 0.0961 -0.0919 0.0944 1 D/E -0.1240 0.0353 0.3141 -0.1560 0.0120 1 Panel C AB-to-TA CP-to-TA ROA M/B D/E AB-to-TA 1 CP-to-TA 0.1991 1 ROA 0.1879 -0.0778 1 M/B 0.0278 0.2189 0.0944 1 D/E -0.1863 -0.0948 -0.1600 0.0120 1 Notes: 1. data include both permanent CEOs firms and transient CEO firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent
pay, AB-to-TP denotes annual bonus as a percentage of total pay, CP-to-TP denotes contingent pay as a percentage of total pay, AB-to-TA denotes annual bonus as a percentage of total assets, CP-to-TA denotes contingent pay as a percentage of total assets, TP-to-TA denotes total pay as a percentage of total assets, ln(ta) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market-to-book ratio, D/E denotes debt to equity ratio.
70
Table 8.2: Correlation of variables in the family-controlled firms Panel A ln (AB) ln (CP) ln (TA) ROA M/B D/E ln(AB) 1.0000
ln(CP) 0.0362 1.0000 ln(TA) -0.0612 0.1276 1.0000
ROA 0.4371 -0.2052 -0.2257 1.0000 M/B 0.1011 -0.0217 0.0811 0.2520 1.0000
D/E -0.1287 0.0305 -0.0428 -0.2187 0.3237 1.0000 Panel B AB-to-TP CP-to-TP ln(TA) ROA M/B D/E AB-to-TP 1.0000
CP-to-TP -0.6758 1.0000 ln(TA) -0.1354 0.1640 1.0000
ROA 0.3638 -0.3199 -0.2257 1.0000 M/B -0.0413 0.0477 0.0811 0.2520 1.0000
D/E -0.2121 0.0104 -0.0428 -0.2187 0.3237 1.0000 Panel C AB-to-TA CP-to-TA ROA M/B D/E AB-to-TA 1.0000
CP-to-TA 0.0885 1.0000 ROA 0.2971 -0.2699 1.0000
M/B -0.0725 0.0061 0.2520 1.0000 D/E -0.0360 0.1140 -0.2187 0.3237 1.0000
Notes: 1. data include family controlled firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent pay, AB-to-TP denotes annual bonus as a percentage of total pay, CP-to-TP denotes contingent pay as a percentage of total pay, AB-to-TA denotes annual bonus as a percentage of total assets, CP-to-TA denotes contingent pay as a percentage of total assets, TP-to-TA denotes total pay as a percentage of total assets, ln(TA) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market-to-book ratio, D/E denotes debt to equity ratio.
71
Table 8.3: Correlation of variables in the institution-controlled firms Panel A ln (AB) ln (CP) ln (TA) ROA M/B D/E ln(AB) 1.0000
ln(CP) 0.1905 1.0000 ln(TA) -0.0238 0.0098 1.0000
ROA 0.3605 0.0363 0.0872 1.0000 M/B 0.1107 0.1095 -0.1661 0.0600 1.0000
D/E -0.1242 0.0088 0.2027 -0.0283 0.0926 1.0000 Panel B AB-to-TP CP-to-TP ln(TA) ROA M/B D/E AB-to-TP 1.0000
CP-to-TP -0.6347 1.0000 ln(TA) 0.0197 0.0673 1.0000
ROA 0.1778 -0.0066 0.0872 1.0000 M/B -0.0187 0.1341 -0.1661 0.0600 1.0000
D/E -0.1963 0.0904 0.2027 -0.0283 0.0926 1.0000 Panel C AB-to-TA CP-to-TA ROA M/B D/E AB-to-TA 1.0000
CP-to-TA 0.1577 1.0000 ROA 0.2560 -0.1152 1.0000
M/B 0.0418 0.2539 0.0600 1.0000 D/E -0.2564 0.0505 -0.0283 0.0926 1.0000
Notes: 1. data include institution controlled firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent pay, AB-to-TP denotes annual bonus as a percentage of total pay, CP-to-TP denotes contingent pay as a percentage of total pay, AB-to-TA denotes annual bonus as a percentage of total assets, CP-to-TA denotes contingent pay as a percentage of total assets, TP-to-TA denotes total pay as a percentage of total assets, ln(TA) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market-to-book ratio, D/E denotes debt to equity ratio.
72
Table 8.4: Correlation of variables in the widely-held firms Panel A ln (AB) ln (CP) ln (TA) ROA M/B D/E ln(AB) 1.0000
ln(CP) 0.3095 1.0000 ln(TA) 0.3199 0.3782 1.0000
ROA 0.0750 -0.0847 -0.0933 1.0000 M/B -0.0106 0.0424 -0.0211 0.0750 1.0000
D/E 0.0066 0.1376 0.5780 -0.2474 -0.1536 1.0000 Panel B AB-to-TP CP-to-TP ln(TA) ROA M/B D/E AB-to-TP 1.0000
CP-to-TP -0.5078 1.0000 ln(TA) -0.0250 0.2611 1.0000
ROA 0.1581 -0.0645 -0.0933 1.0000 M/B -0.0321 0.0758 -0.0211 0.0750 1.0000
D/E -0.0528 0.0419 0.5780 -0.2474 -0.1536 1.0000 Panel C AB-to-TA CP-to-TA ROA M/B D/E AB-to-TA 1.0000
CP-to-TA 0.2269 1.0000 ROA 0.1233 -0.0635 1.0000
M/B 0.0054 0.1709 0.0750 1.0000 D/E -0.2504 -0.2474 -0.2474 -0.1536 1.0000
Notes: 1. data include widely-held firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent pay, AB-to-TP denotes annual bonus as a percentage of total pay, CP-to-TP denotes contingent pay as a percentage of total pay, AB-to-TA denotes annual bonus as a percentage of total assets, CP-to-TA denotes contingent pay as a percentage of total assets, TP-to-TA denotes total pay as a percentage of total assets, ln(TA) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market-to-book ratio, D/E denotes debt to equity ratio.
73
Table 9.1: The impact of ownership structure on the natural log of annual bonus (OLS) In model1, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 T-stat 2 T-stat 3 T-stat 4 T-stat α -1.6501 -0.5591 -3.3069 -1.0219 -3.3017 -1.0202 -18.5654 -4.4370*** Ln(TA) 0.5147 3.7519*** 0.5964 4.0201*** 0.6009 4.0652*** 1.4043 7.2911*** ROA 0.1576 5.3582*** 0.1538 5.0025*** 0.1505 4.9438*** 0.0260 0.5016 M/B 0.1318 1.0031 0.1212 0.9246 0.1078 0.8259 -0.2208 -0.9994 D/E -0.0033 -1.6370 -0.0032 -1.5868 -0.0032 -1.5858 -0.0122 -4.2756*** Dyr04 0.7641 1.1134 0.7476 1.0848 0.7997 1.1543 0.9134 1.3562 Dyr05 1.2122 1.8491* 1.1897 1.8070* 1.3130 1.9933** 1.6056 2.5129** Dyr06 1.2876 1.9792** 1.2503 1.8996* 1.2995 1.9691** 1.4806 2.3589** Dyr07 1.0417 1.5481 0.9905 1.4625 0.9804 1.4440 1.3436 2.0263** Dfin -0.6837 -0.9580 -0.7181 -1.0061 -1.5933 -2.0838** Degy -0.2428 -0.4346 -0.2888 -0.5144 -0.6361 -1.1167 Dmat 0.4733 0.9613 0.4779 0.9708 -0.7076 -1.3563 D1yc -1.2112 -1.3262 -1.2610 -1.4504 Dcon 28.9372 4.4918*** Dcon*ln(TA) -1.4380 -4.8760*** Dcon*ROA 0.2053 3.2728*** Dcon*M/B 0.4452 1.6966* Dcon*D/E 0.0088 2.4625** Adjusted R-square 0.087 0.086 0.088 0.150 F-statistic 7.614 5.762 5.481 6.822 Observations 560 560 560 560
74
Table 9.2: Natural log of annual bonus in widely-held, institution-controlled, and family-controlled firms (OLS) In model 5, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). Moreover, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 T-stat 6 T-stat α -19.7558 -4.6390*** 11.9361 1.9702** Ln(TA) 1.4552 7.4315*** -0.0972 -0.3482 ROA 0.0241 0.4649 0.2009 5.4715*** M/B -0.2094 -0.9423 0.2223 1.5724 D/E -0.0121 -4.2471*** -0.0088 -1.8478* Dyr04 0.9867 1.4704 0.9867 1.4704 Dyr05 1.5775 2.4698** 1.5775 2.4698** Dyr06 1.4743 2.3639** 1.4743 2.3639** Dyr07 1.3295 2.0242** 1.3295 2.0242** Dfin -1.8389 -2.3194** -1.8389 -2.3194** Degy -0.2803 -0.4880 -0.2803 -0.4880 Dmat -0.5409 -0.9991 -0.5409 -0.9991 D1yc -1.2526 -1.4444 -1.2526 -1.4444 DF 22.3362 2.3589** -9.3557 -0.8522 DF*ln(TA) -1.1513 -2.6934** 0.4011 0.8051 DF*ROA 0.4937 4.2209*** 0.3169 2.8597*** DF*M/B 0.2789 0.4921 -0.1528 -0.2790 DF*D/E 0.0105 2.6905*** 0.0072 1.3207 DI 31.6919 4.0301*** DI*ln(TA) -1.5524 -4.2574*** DI*ROA 0.1768 2.7932*** DI*M/B 0.4316 1.6151 DI*D/E 0.0033 0.6113 Adjusted R-square 0.159 0.159 F-statistic 5.799 5.799 Observations 560 560
75
Table 10.1: The impact of ownership structure on the natural log of contingent compensation (OLS) In model1, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 T-stat 2 T-stat 3 T-stat 4 T-stat α -4.3253 -1.6114 -13.568 -3.8456*** -13.5724 -3.8310*** -26.7482 -6.0524*** Ln(TA) 0.6792 5.4893*** 1.1033 6.7929*** 1.0995 6.7473*** 1.7452 8.5551*** ROA -0.0315 -1.0019 -0.0656 -1.9987** -0.0629 -1.9135* -0.0859 -1.9027* M/B 0.2697 1.9932** 0.2556 1.8801* 0.2667 1.9627* 0.1444 0.5971 D/E -0.0007 -0.3462 -0.0011 -0.5287 -0.0011 -0.5303 -0.0059 -1.9502* Dyr04 0.8207 1.0619 0.7383 0.9760 0.6948 0.9187 0.7988 1.0679 Dyr05 1.3577 1.8249* 1.2298 1.7011* 1.1270 1.5434 1.2912 1.7714* Dyr06 0.6171 0.7877 0.4328 0.5684 0.3917 0.5142 0.5374 0.7057 Dyr07 0.9187 1.1794 0.6707 0.8690 0.6791 0.8780 0.9242 1.2006 Dfin -2.2535 -2.6250*** -2.2248 -2.5719** -2.9027 -3.2105*** Degy 1.8168 2.8715*** 1.8551 2.9257*** 1.7478 2.7077** Dmat 2.1719 3.8515*** 2.1681 3.8380*** 1.3828 2.3468** D1yc 1.0094 1.2034 0.9620 1.1756 Dcon 25.8375 4.3670*** Dcon*ln(TA) -1.2128 -4.4807*** Dcon*ROA 0.0359 0.5624 Dcon*M/B 0.1252 0.4352 Dcon*D/E 0.0043 1.0701 Adjusted R-squared 0.043 0.088 0.088 0.108 F-statistic 4.152 5.927 5.539 4.995 Observations 560 560 560 560
76
Table 10.2: Natural log of contingent compensation in widely-held, institution-controlled, and family-controlled firms (OLS) In model 5, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). Moreover, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 T-stat 6 T-stat α -26.9059 -6.0247*** 3.0363 0.5229 Ln(TA) 1.7523 8.5198*** 0.3571 1.3103 ROA -0.0871 -1.9162* -0.0224 -0.4271 M/B 0.1386 0.5689 0.2433 1.4815 D/E -0.0061 -1.9752** -0.0044 -1.0231 Dyr04 0.7966 1.0531 0.7966 1.0531 Dyr05 1.3935 1.9001* 1.3935 1.9001 Dyr06 0.6226 0.8141 0.6226 0.8141 Dyr07 1.0040 1.2973 1.0040 1.2973 Dfin -2.9082 -3.2275*** -2.9082 -3.2275*** Degy 1.8000 2.6277*** 1.8000 2.6277*** Dmat 1.2242 2.0911** 1.2242 2.0911** D1yc 0.8280 1.0160 0.8280 1.0160 DF 18.7768 2.1508** -11.1654 -1.1608 DF*ln(TA) -0.9050 -2.3270** 0.4902 1.1330 DF*ROA -0.1455 -1.1179 -0.2102 -1.5696 DF*M/B 0.2823 0.5404 0.1776 0.3602 DF*D/E 0.0044 0.8960 0.0028 0.4853 DI 29.9422 4.2630*** DI*ln(TA) -1.3951 -4.2550*** DI*ROA 0.0647 0.9386 DI*M/B 0.1046 0.3532 DI*D/E 0.0016 0.2972 Adjusted R-squared 0.108 0.108 F-statistic 4.076 4.076 Observations 560 560
77
Table 11.1: The impact of ownership structure on the annual bonus as a proportion of total assets (Tobit) In model1, use return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 T-stat 2 T-stat 3 T-stat 4 T-stat α 0.0002 2.5188** 0.0002 3.0105*** 0.0002 3.1780*** 0.0005 4.1727*** ROA 0.00002 3.6438*** 0.00002 3.4572*** 0.00002 3.3909*** 0.000008 1.0858 M/B 0.000009 0.5368 0.000006 0.3944 0.000005 0.2950 -0.00003 -0.9553 D/E -0.000001 -2.9567*** -0.000001 -2.8902*** -0.000001 -2.9014*** -0.000002 -4.7242*** Dyr04 0.0001 1.2307 0.0001 1.2168 0.0001 1.2836 0.00009 1.1885 Dyr05 0.0001 1.7587* 0.0001 1.7671* 0.0002 1.9174* 0.0002 2.0230** Dyr06 0.00008 1.1121 0.00008 1.1129 0.0001 1.1806 0.0001 1.2063 Dyr07 0.0001 1.6496* 0.0001 1.6536* 0.0001 1.6172 0.0001 1.7764* Dfin -0.0002 -2.6492*** -0.0002 -2.6927** -0.0001 -1.8260* Degy -0.0001 -0.8092 -0.00007 -0.8914 -0.0001 -1.4529 Dmat -0.000004 -0.0658 -0.000004 -0.0620 -0.000002 -0.0238 D1yc -0.0002 -1.9539* -0.0002 -1.9446* Dcon -0.0004 -3.0213*** Dcon*ROA 0.00002 1.6394 Dcon*M/B 0.00005 1.4442 Dcon*D/E 0.000001 2.6616*** Adjusted R-square 0.052 0.056 0.058 0.075 Akaike info criterion -9.802 -9.802 -9.803 -9.813 Schwarz criterion -9.733 -9.709 -9.702 -9.682 Observations 560 560 560 560
78
Table 11.2: Annual bonus as a proportion of total assets in widely-held, institution-controlled, and family-controlled firms (Tobit) In model 5, use return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). We add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family-controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 Z-stat 6 Z-stat α 0.0005 4.1301*** 0.0002 2.4275** ROA 0.000008 1.0944 0.00002 2.8867*** M/B -0.00003 -0.9758 0.00002 1.2283 D/E -0.000002 -4.8765*** -0.000002 -3.6236*** Dyr04 0.0001 1.2943 0.0001 1.2943 Dyr05 0.0002 2.1752** 0.0002 2.1752** Dyr06 0.0001 1.3939 0.0001 1.3939 Dyr07 0.0001 1.9003* 0.0001 1.9003* Dfin -0.00009 -1.2541 -0.00009 -1.2541 Degy -0.00008 -1.0340 -0.00008 -1.0340 Dmat -0.00002 -0.2825 -0.00002 -0.2825 D1yc -0.0002 -2.3242** -0.0002 -2.3242** DF -0.0005 -3.1721*** -0.0002 -1.6692* DF*ROA 0.00004 2.7022*** 0.00003 1.9143* DF*M/B -0.00003 -0.6156 -0.00008 -1.8212* DF*D/E 0.000002 4.2951*** 0.000002 3.5371*** DI -0.0003 -1.9404* DI*ROA 0.00001 1.2409 DI*M/B 0.00005 1.4602 DI*D/E 0.0000001 -0.1166 Adjusted R-square 0.085 0.085 Akaike info criterion -9.822 -9.822 Schwarz criterion -9.660 -9.660 Observations 560 560
79
Table 12.1: The impact of ownership structure on the contingent compensation as a proportion of total assets (Tobit) In model1, use return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 Z-stat 2 Z-stat 3 Z-stat 4 Z-stat α 0.0006 2.0627** 0.0004 1.5622 0.0004 1.3873 0.0006 1.7319* ROA -0.00004 -2.2782** -0.00005 -3.0313*** -0.00005 -2.9291*** -0.00005 -1.8982* M/B 0.0003 4.1542*** 0.0003 4.2522*** 0.0003 4.3373*** 0.0002 2.3265** D/E 0.0000 -3.1833*** -0.000001 -2.2295** 0.00000 -2.2581** 0.0000 -2.5873** Dyr04 -0.0002 -0.4758 -0.0002 -0.4857 -0.0002 -0.5739 -0.0002 -0.5519 Dyr05 -0.0001 -0.3165 -0.0001 -0.2941 -0.0002 -0.4981 -0.0002 -0.4436 Dyr06 -0.0006 -1.7066* -0.0005 -1.6866* -0.0006 -1.7607* -0.0006 -1.7169* Dyr07 -0.0006 -1.8566* -0.0006 -1.8565* -0.0006 -1.8633* -0.0005 -1.7501* Dfin -0.0006 -2.6783*** -0.0006 -2.5737** -0.0005 -1.9563* Degy 0.0008 2.5131** 0.0008 2.6140*** 0.0008 2.5224** Dmat 0.0008 2.9579*** 0.0008 2.9890*** 0.0008 2.8227*** D1yc 0.0007 1.5803 0.0007 1.5993 Dcon -0.0004 -0.9990 Dcon*ROA -0.000003 -0.1018 Dcon*M/B 0.0001 0.6395 Dcon*D/E 0.000002 1.2832 Adjusted R-square 0.051 0.088 0.095 0.093 Akaike info criterion -7.416 -7.454 -7.458 -7.446 Schwarz criterion -7.347 -7.361 -7.357 -7.315 Observations 560 560 560 560
80
Table 12.2: Contingent compensation as a proportion of total assets in widely-held, institution-controlled, and family-controlled firms (Tobit) In model 5, use return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). We add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family-controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 Z-stat 6 Z-stat α 0.0007 1.8813* 0.0004 1.0006 ROA 0.0000 -1.8756* -0.00005 -2.1724** M/B 0.0002 2.2660** 0.0003 3.4936*** D/E -0.000003 -2.8055*** 0.0000002 0.0765 Dyr04 -0.0002 -0.5644 -0.0002 -0.5644 Dyr05 -0.0002 -0.4225 -0.0002 -0.4225 Dyr06 -0.0005 -1.6851* -0.0005 -1.6851* Dyr07 -0.0005 -1.6869* -0.0005 -1.6869* Dfin -0.0004 -1.5971 -0.0004 -1.5971 Degy 0.0007 2.0223** 0.0007 2.0223** Dmat 0.0007 2.4898** 0.0007 2.4898** D1yc 0.0007 1.5850 0.0007 1.5850 DF -0.0004 -1.2073 -0.0002 -0.4803 DF*ROA 0.0000 -0.4345 -0.00001 -0.3104 DF*M/B -0.0001 -0.8188 -0.0002 -1.5568 DF*D/E 0.000003 2.1373** -0.0000003 -0.0961 DI -0.0003 -0.5942 DI*ROA -0.000004 -0.1323 DI*M/B 0.0001 0.6431 DI*D/E 0.000003 0.8801 Adjusted R-square 0.095 0.095 Akaike info criterion -7.443 -7.443 Schwarz criterion -7.281 -7.281 Observations 560 560
81
Table 13.1: The impact of ownership structure on the annual bonus as a proportion of total pay (Tobit) In model1, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% level. Independent Variable Model 1 T-stat 2 T-stat 3 T-stat 4 T-stat α 0.0905 0.7460 0.3643 2.3812 ** 0.3700 2.4166** 0.2155 1.0415 Ln(TA) 0.0032 0.5911 -0.0093 -1.3424 -0.0092 -1.3249 -0.0007 -0.0684 ROA 0.0063 4.7976*** 0.0074 5.4197 *** 0.0071 5.2828 *** 0.0049 2.1817 ** M/B -0.0045 -0.7842 -0.0047 -0.8131 -0.0057 -0.9953 -0.0129 -1.4743 D/E -0.0002 -2.0458 ** -0.0002 -1.9007 * -0.0002 -1.9165 * -0.0002 -1.2373 Dyr04 0.0240 0.8049 0.0263 0.8986 0.0298 1.0196 0.0324 1.0907 Dyr05 0.0432 1.4574 0.0476 1.6387 0.0567 1.9388 * 0.0622 2.0955 ** Dyr06 0.0760 2.4716 ** 0.0819 2.7195 *** 0.0852 2.8249 *** 0.0875 2.8999 *** Dyr07 0.0624 2.0556 ** 0.0698 2.3438 ** 0.0685 2.3052 ** 0.0735 2.4436 ** Dfin 0.0636 1.7578 * 0.0613 1.6849 * 0.0463 1.1563 Degy -0.0712 -2.7102 *** -0.0750 -2.8698 *** -0.0811 -3.0798 *** Dmat -0.0425 -1.9338 * -0.0423 -1.9413 * -0.0625 -2.6141 *** D1yc -0.0941 -2.8782 *** -0.0965 -2.9541 *** Dcon 0.3180 1.1278 Dcon*ln(TA) -0.0163 -1.2575 Dcon*ROA 0.0037 1.3641 Dcon*M/B 0.0118 1.0327 Dcon*D/E 0.0000 0.1033 Adjusted R-square 0.045 0.075 0.083 0.076 Akaike info criterion 0.159 0.143 0.135 0.143 Schwarz criterion 0.236 0.244 0.243 0.299 Observations 560 560 560 560
82
Table 13.2: Annual bonus as a proportion of total pay in widely-held, institution-controlled, and family-controlled firms (Tobit) In model 5, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). We add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family-controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 Z-stat 6 Z-stat α 0.1923 0.9250 0.4096 1.4503 Ln(TA) 0.0002 0.0243 -0.0113 -0.8612 ROA 0.0048 2.1770 ** 0.0072 4.2559 *** M/B -0.0124 -1.4140 0.0032 0.3779 D/E -0.0002 -1.2015 -0.0005 -2.2467 ** Dyr04 0.0339 1.1379 0.0339 1.1379 Dyr05 0.0599 2.0119 ** 0.0599 2.0119 ** Dyr06 0.0857 2.8717 *** 0.0857 2.8717 *** Dyr07 0.0710 2.3805 ** 0.0710 2.3805 ** Dfin 0.0435 1.0688 0.0435 1.0688 Degy -0.0671 -2.5134 ** -0.0671 -2.5134 ** Dmat -0.0536 -2.2007 ** -0.0536 -2.2007 ** D1yc -0.0986 -2.9901 *** -0.0986 -2.9901 *** DF 0.3970 0.9982 0.1797 0.3883 DF*ln(TA) -0.0184 -1.0465 -0.0069 -0.3359 DF*ROA 0.0178 2.9426 *** 0.0155 2.6423 *** DF*M/B -0.0134 -0.6925 -0.0290 -1.4962 DF*D/E 0.0001 0.6584 0.0004 1.6909 DI 0.2173 0.6146 DI*ln(TA) -0.0115 -0.6970 DI*ROA 0.0023 0.8452 DI*M/B 0.0156 1.2706 DI*D/E -0.0003 -1.0204 Adjusted R-square 0.083 0.083 Akaike info criterion 0.140 0.140 Schwarz criterion 0.326 0.326 Observations 560 0.083
83
Table 14.1: The impact of ownership structure on the contingent compensation as a proportion of total pay (Tobit) In model1, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 Z-stat 2 Z-stat 3 Z-stat 4 Z-stat α -0.4562 -2.6005*** -1.1850 -5.4820*** -1.1814 -5.4880*** -1.6976 -5.9681*** Ln(TA) 0.0381 4.8804*** 0.0711 7.2053*** 0.0702 7.1432*** 0.0953 7.3733*** ROA -0.0032 -1.5541 -0.0062 -2.9890*** -0.0056 -2.7216*** -0.0052 -1.4761 M/B 0.0233 2.6201*** 0.0233 2.6037*** 0.0254 2.8972*** 0.0200 1.3787 D/E -0.0001 -0.7813 -0.0001 -1.1694 -0.0001 -1.2016 -0.0004 -2.5275** Dyr04 0.0254 0.5131 0.0188 0.3917 0.0100 0.2115 0.0144 0.3069 Dyr05 0.0389 0.8017 0.0277 0.5984 0.0066 0.1437 0.0129 0.2775 Dyr06 -0.0258 -0.5159 -0.0406 -0.8612 -0.0490 -1.0503 -0.0427 -0.9081 Dyr07 -0.0173 -0.3562 -0.0364 -0.7680 -0.0349 -0.7399 -0.0242 -0.5094 Dfin -0.1570 -3.2061*** -0.1501 -3.0390*** -0.1686 -3.1904*** Degy 0.1953 4.5852*** 0.2026 4.8187*** 0.2011 4.7260*** Dmat 0.1546 4.1607*** 0.1538 4.1914*** 0.1263 3.2045*** D1yc 0.1996 3.5715*** 0.1976 3.5261*** Dcon 0.9688 2.5093** Dcon*ln(TA) -0.0452 -2.5691** Dcon*ROA -0.0009 -0.2169 Dcon*M/B 0.0058 0.3149 Dcon*D/E 0.0003 1.4012 Adjusted R-square 0.049 0.109 0.134 0.132 Akaike info criterion 1.035 0.962 0.942 0.949 Schwarz criterion 1.112 1.062 1.051 1.096 Observations 560 560 560 560
84
Table 14.2: Contingent compensation as a proportion of total pay in widely-held, institution-controlled, and family-controlled firms (Tobit) In model 5, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). We add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family-controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 Z-stat 6 Z-stat α -1.6953 -5.9276*** -0.5412 -1.4740 Ln(TA) 0.0953 7.3434*** 0.0419 2.4206** ROA -0.0053 -1.4947 -0.0040 -1.3962 M/B 0.0195 1.3447 0.0205 1.6673* D/E -0.0004 -2.5667** -0.0001 -0.4769 Dyr04 0.0112 0.2363 0.0112 0.2363 Dyr05 0.0183 0.3939 0.0183 0.3939 Dyr06 -0.0394 -0.8419 -0.0394 -0.8419 Dyr07 -0.0226 -0.4750 -0.0226 -0.4750 Dfin -0.1699 -3.1934*** -0.1699 -3.1934*** Degy 0.1978 4.4247*** 0.1978 4.4247*** Dmat 0.1176 2.9843*** 0.1176 2.9843*** D1yc 0.1908 3.4065*** 0.1908 3.4065*** DF 0.9299 1.6910* -0.2242 -0.3649 DF*ln(TA) -0.0454 -1.8756* 0.0079 0.2889 DF*ROA -0.0197 -2.4735** -0.0209 -2.6942*** DF*M/B 0.0480 1.5943 0.0470 1.5943 DF*D/E 0.0002 0.7554 -0.0001 -0.2872 DI 1.1541 2.4893** DI*ln(TA) -0.0534 -2.4692** DI*ROA 0.0013 0.2805 DI*M/B 0.0010 0.0519 DI*D/E 0.0003 0.8486 Adjusted R-square 0.136 0.136 Akaike info criterion 0.954 0.954 Schwarz criterion 1.140 1.140 Hannan-Quinn criter. 1.027 1.027 Observations 560 560
85
Table 15: Correlation of variables in the family-controlled, institution-controlled, and widely-held firms Panel A ln (AB) ln (CP) ln (ta) ROA M/B D/E TMR ln(AB) 1 ln(CP) 0.2139 1 ln(ta) 0.1428 0.2056 1 ROA 0.2371 -0.0419 -0.0642 1 M/B 0.0547 0.0629 -0.0919 0.0944 1 D/E -0.0687 0.0563 0.3141 -0.1600 0.0120 1 TMR 0.0351 0.0191 -0.1706 0.1263 0.2509 -0.0873 1 Panel B AB-to-TP CP-to-TP ln(ta) ROA M/B D/E TMR AB-to-TP 1 CP-to-TP -0.5885 1 ln(ta) -0.0167 0.1723 1 ROA 0.1830 -0.0651 -0.0642 1 M/B -0.0312 0.0961 -0.0919 0.0944 1 D/E -0.1240 0.0353 0.3141 -0.1560 0.0120 1 TMR 0.0323 0.0240 -0.1706 0.1263 0.2509 -0.0873 1 Panel C AB-to-TA CP-to-TA ROA M/B D/E TMR AB-to-TA 1 CP-to-TA 0.1991 1 ROA 0.1879 -0.0778 1 M/B 0.0278 0.2189 0.0944 1 D/E -0.1863 -0.0948 -0.1600 0.0120 1 TMR 0.1495 0.1984 0.1263 0.2509 -0.0873 1 Notes: 1. data include widely-held firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent pay, ln(TA) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market to book ratio, TMR denotes total market return ratio, and D/E denotes debt to equity ratio.
86
Table 16.1: The impact of ownership structure on the natural log of annual bonus (OLS) In model9, use total assets (ln (TA)), return on assets (ROA), total market return (TMR), debt to equity ratio (d/e), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 10, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 11, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 12, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 9 T-stat 10 T-stat 11 T-stat 12 T-stat Α -1.5454 -0.5130 -3.3902 -1.0523 -3.4299 -1.0652 -19.6036 -4.6953*** Ln(TA) 0.5168 3.7199*** 0.6051 4.0535*** 0.6106 4.1064*** 1.4112 7.3240*** ROA 0.1566 5.2311*** 0.1521 4.8853*** 0.1486 4.8407*** 0.0205 0.3876 TMR 0.4111 1.0525 0.5052 1.2611 0.4878 1.2181 0.6854 1.3182 D/E -0.0032 -1.5048 -0.0031 -1.4458 -0.0031 -1.4582 -0.0112 -3.9489*** Dyr04 0.8465 1.2155 0.8508 1.2178 0.9009 1.2823 0.9274 1.3685 Dyr05 1.2694 1.9345* 1.2448 1.8939* 1.3654 2.0735** 1.4945 2.3406** Dyr06 1.4263 2.1523** 1.4076 2.0981** 1.4497 2.1546** 1.5865 2.4680** Dyr07 1.1932 1.7440* 1.1692 1.7017* 1.1513 1.6713* 1.3898 2.0571** Dfin -0.7448 -1.0436 -0.7786 -1.0908 -1.6228 -2.1277** Degy -0.3771 -0.6636 -0.4177 -0.7326 -0.6913 -1.1919 Dmat 0.5131 1.0361 0.5134 1.0366 -0.5893 -1.1202 D1yc -1.2368 -1.3439 -1.2003 -1.3737 Dcon 31.1011 4.9057*** Dcon*ln(TA) -1.4748 -5.0489*** Dcon*ROA 0.2142 3.3102*** Dcon*TMR -0.5022 -0.6543 Dcon*D/E 0.0082 2.2521** Adjusted R-square 0.086 0.087 0.089 0.148 F-statistic 7.597 5.821 5.545 6.717 Observations 560 560 560 560
87
Table 16.2: Natural log of annual bonus in widely-held, institution-controlled, and family-controlled firms (OLS) In model 13, use total assets (ln (TA)), return on assets (ROA), total market return (TMR), debt to equity ratio (d/e), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). Moreover, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 13 T-stat 14 T-stat α -20.8925 -4.9475*** 13.5344 2.2707** Ln(TA) 1.4690 7.5444*** -0.1466 -0.5300 ROA 0.0187 0.3553 0.2037 5.1417*** TMR 0.6552 1.2639 -0.0527 -0.0870 D/E -0.0111 -3.9212*** -0.0078 -1.6359 Dyr04 0.9991 1.4815 0.9991 1.4815 Dyr05 1.4806 2.3224** 1.4806 2.3224** Dyr06 1.5273 2.3941** 1.5273 2.3941** Dyr07 1.3596 2.0257** 1.3596 2.0257** Dfin -1.9181 -2.4531** -1.9181 -2.4531** Degy -0.3204 -0.5484 -0.3204 -0.5484 Dmat -0.3792 -0.6973 -0.3792 -0.6973 D1yc -1.1821 -1.3776 -1.1821 -1.3776 DF 24.4537 2.5635** -9.9732 -0.9122 DF*ln(TA) -1.2197 -2.8614*** 0.3960 0.8000 DF*ROA 0.4749 4.4327*** 0.2899 2.8530*** DF*M/B 2.1278 0.9085 2.8356 1.2057 DF*D/E 0.0094 2.5068** 0.0061 1.1347 DI 34.4269 4.4862*** DI*ln(TA) -1.6156 -4.5090*** DI*ROA 0.1850 2.8174*** DI*M/B -0.7079 -0.9026 DI*D/E 0.0033 0.6055 Adjusted R-square 0.159 0.159 F-statistic 5.818 5.818 Observations 560 560
88
Table 17.1: The impact of ownership structure on the natural log of contingent pay (OLS) In model9, use total assets (ln (TA)), return on assets (ROA), total market return (TMR), debt to equity ratio (d/e), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 10, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 11, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 12, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 9 T-stat 10 T-stat 11 T-stat 12 T-stat α -3.9256 -1.4493 -13.292 -3.908*** -13.2629 -3.8871*** -27.4721 -6.1060*** Ln(TA) 0.6777 5.3508*** 1.106 6.940*** 1.1022 6.8942*** 1.7839 8.5514*** ROA -0.0321 -1.0074 -0.066 -1.989** -0.0637 -1.9071* -0.0894 -1.9506* TMR 0.6987 1.6289 0.738 1.695* 0.7505 1.7213* 0.6358 1.2311 D/E -0.0004 -0.2007 -0.001 -0.350 -0.0008 -0.3449 -0.0061 -1.9478* Dyr04 0.9582 1.2148 0.885 1.145 0.8475 1.0964 0.9637 1.2621 Dyr05 1.4713 1.9590* 1.339 1.838* 1.2494 1.6954* 1.3972 1.9218* Dyr06 0.8659 1.0956 0.686 0.892 0.6543 0.8508 0.7993 1.0399 Dyr07 1.1827 1.4853 0.943 1.196 0.9562 1.2099 1.2284 1.5348 Dfin -2.336 -2.752*** -2.3106 -2.7028*** -3.0148 -3.3609*** Degy 1.606 2.502** 1.6366 2.5447** 1.5676 2.4021** Dmat 2.254 3.995*** 2.2541 3.9838*** 1.4669 2.5027** D1yc 0.9196 1.0860 0.8981 1.0910 Dcon 27.0490 4.5100*** Dcon*ln(TA) -1.2608 -4.5493*** Dcon*ROA 0.0353 0.5407 Dcon*TMR 0.4093 0.4802 Dcon*D/E 0.0049 1.1771 Adjusted R-square 0.041 0.087 0.087 0.109 F-statistic 3.958 5.837 5.437 5.038 Observations 560 560 560 560
89
Table 17.2: Natural log of contingent pay in widely-held, institution-controlled, and family-controlled firms (OLS) In model 13, use total assets (ln (TA)), return on assets (ROA), total market return (TMR), debt to equity ratio (d/e), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). Moreover, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 13 T-stat 14 T-stat α -27.5014 -6.0868*** 3.7031 0.6546 Ln(TA) 1.7852 8.5409*** 0.3380 1.2566 ROA -0.0901 -1.9578* -0.0270 -0.5047 TMR 0.6211 1.1990 0.8498 1.1897 D/E -0.0062 -1.9695** -0.0032 -0.7399 Dyr04 0.9442 1.2255 0.9442 1.2255 Dyr05 1.4785 2.0200** 1.4785 2.0200** Dyr06 0.8559 1.1061 0.8559 1.1061 Dyr07 1.2802 1.6018 1.2802 1.6018 Dfin -2.9918 -3.3859*** -2.9918 -3.3859*** Degy 1.6006 2.3313** 1.6006 2.3313** Dmat 1.3042 2.1974** 1.3042 2.1974** D1yc 0.7819 0.9557 0.7819 0.9557 DF 19.1567 2.0714** -12.0478 -1.1966 DF*ln(TA) -0.9084 -2.1897** 0.5388 1.1725 DF*ROA -0.1157 -0.7787 -0.1788 -1.1791 DF*M/B 0.5961 0.1725 0.3673 0.1060 DF*D/E 0.0053 1.0954 0.0022 0.3961 DI 31.2045 4.4064*** DI*ln(TA) -1.4473 -4.3301*** DI*ROA 0.0631 0.8968 DI*M/B 0.2287 0.2635 DI*D/E 0.0031 0.5622 Adjusted R-square 0.108 0.108 F-statistic 4.079 4.079 Observations 560 560
90
Appendix II: Figures Figure 1.1: Salary in monetary terms paid by permanent CEO firms: comparison between
widely-held, institution-controlled, and family controlled firms (Adjusted for inflation)
Figure1.2: Bonus in monetary terms paid by permanent CEO firms: comparison between widely-
held, institution-controlled, and family controlled firms
0
100000
200000
300000
400000
500000
600000
700000
800000
900000
1000000
2003 2004 2005 2006 2007
Family
Institution
Widely Held
0
500000
1000000
1500000
2000000
2500000
3000000
2003 2004 2005 2006 2007
Family
Institution
Widely Held
91
Figure1.3: Contingent compensation in monetary terms paid by permanent CEO firms:
comparison between widely-held, institution-controlled, and family controlled firms
Figure1.4: Total compensation in monetary terms paid by permanent CEO firms: comparison
between widely-held, institution-controlled, and family controlled firms
0
500000
1000000
1500000
2000000
2500000
3000000
3500000
4000000
4500000
5000000
2003 2004 2005 2006 2007
Family
Institution
Widely Held
0
1000000
2000000
3000000
4000000
5000000
6000000
7000000
8000000
9000000
2003 2004 2005 2006 2007
Family
Institution
Widely Held
92
Figure 2.1: Salary as a percentage of total compensation paid by permanent CEO firms:
comparison between widely-held, institution-controlled, and family controlled firms
Figure 2.2: Annual Bonus as a percentage of total compensation paid by permanent CEO firms:
comparison between widely-held, institution-controlled, and family controlled firms
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
2003 2004 2005 2006 2007
Family Bonus
Institution Bonus
Widely-held Bonus
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
2003 2004 2005 2006 2007
Family Salary
Institution Salary
Widely-Held Salary
93
Figure 2.3: Contingent Compensation as a percentage of total compensation paid by permanent
CEO firms: comparison between widely-held, institution-controlled, and family controlled firms
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
2003 2004 2005 2006 2007
Family Contingent Pay
Institution Contingent Pay
Widely-held Contingent Pay
94
Figure 3.1: Changes in compensation following CEO turnovers in family-controlled firms
Figure 3.2: Changes in compensation following CEO turnovers in institution-controlled firms
0
2000000
4000000
6000000
8000000
10000000
12000000
14000000
salary bonus contingent other total
New
Old
0
500000
1000000
1500000
2000000
2500000
3000000
3500000
4000000
4500000
salary bonus contingent other total
New
Old
95
Figure3.3: Changes in compensation following CEO turnovers in widely-held firms
0
500000
1000000
1500000
2000000
2500000
3000000
3500000
4000000
salary bonus contingent other total
New
Old
96
Figure 4.1: Changes in compensation following CEO retirements in family-controlled firms
Figure 4.2: Changes in compensation following CEO retirements in institution-controlled firms
Notes: retirement firms are firms with paying a large amount of retirement fee to CEOs.
0
2000000
4000000
6000000
8000000
10000000
12000000
14000000
salary bonus contingent other total
New
Old
0
500000
1000000
1500000
2000000
2500000
3000000
3500000
4000000
4500000
salary bonus contingent other total
New
Old
97
Figure 5.1: Total compensation as a function of total assets (all data)
Notes: The horizontal axis is total assets and the vertical axis is total compensation. The unit of the X-axis is $1 billion and the unit of Y-axis is 1 million.
0
10
20
30
40
50
60
70
80
$0 $100 $200 $300 $400 $500 $600 $700
Total Compensation as Function of Total Assets
Total Compensation as Function of Total Assets
98
Figure 5.2: Total compensation as a function of total assets (asset sizes of $20.48 million – 9
billion)
Notes: The horizontal axis is total assets and the vertical axis is total compensation. The unit of the X-axis is $1 billion and the unit of Y-axis is 1 million.
0
2
4
6
8
10
12
14
16
18
20
$0 $1 $2 $3 $4 $5 $6 $7 $8 $9 $10
Total Compensation as Function of Total Assets
Total Compensation as Function of Total Assets
99
Figure 5.3: Total compensation as a function of total assets (asset sizes of $9.1 billion – $55 billion)
Notes: The horizontal axis is total assets and the vertical axis is total compensation. The unit of the X-axis is $1 billion and the unit of Y-axis is 1 million.
0
10
20
30
40
50
60
70
80
$0 $10 $20 $30 $40 $50 $60
Total Compensation as Function of Total Assets
Total Compensation as Function of Total Assets
100
Figure 5.4: Total compensation as a function of total assets (asset sizes larger than $55 billion)
Notes: The horizontal axis is total assets and the vertical axis is total compensation. The unit of the X-axis is $1 billion and the unit of Y-axis is 1 million.
0
5
10
15
20
25
30
$0 $100 $200 $300 $400 $500 $600 $700
Total Compensation as Function of Total Assets
Total Compensation as Function of Total Assets
101
Figure 6.1: Salary and Bonus are as a function of total assets (all data included)
Notes: The horizontal axis is total assets and the vertical axis is salary and bonus. The unit of the X-axis is $1billion and the unit of Y-axis is 1 million.
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0 100 200 300 400 500 600 700
Salary and Bonus as Function of Asset Size
Salary and Bonus as Function of Asset Size
102
Figure 6.2: Salary and Bonus are as a function of total assets (asset sizes of $20.48 million - $55
billion)
Notes: The horizontal axis is total assets and the vertical axis is salary and bonus. The unit of the X-axis is $1billion and the unit of Y-axis is 1 million.
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
0 10 20 30 40 50 60
Salary and Bonus as Function of Asset Size
Salary and Bonus as Function of Asset Size